Investigating the Management of Potentially Cancerous Nonhealing

Investigating the management of potentially cancerous non-healing mouth ulcers in Australian community pharmacies

Brigitte Janse van Rensburg1, Christopher R. Freeman1, Pauline J. Ford2, Meng-Wong Taing1, 1School of Pharmacy, 2School of Dentistry, The University of Queensland, QLD, Australia.

Correspondence:

Dr Meng-Wong Taing, School of Pharmacy, The University of Queensland, Pharmacy Australia Centre of Excellence, 20 Cornwall St, Woolloongabba, QLD 4102, Australia.

Email: [email protected]

Word count: abstract: 249; main text: 3,433 Tables: 4 (2 supplements) Figures: None

Conflicts of interest: None.

Source of Funding

This research that was funded by an Australian Dental Research Fund grant. The sponsors did not have a role in the design of the study, the collection, analysis and interpretation of the data, or in the writing and submission of this manuscript for publication.

Acknowledgments We would like to acknowledge the work of UQ pharmacy student Katelyn Steele with collecting data for this study and the UQ School of Pharmacy, for provision of resources supporting this project. Author Manuscript

This is the author manuscript accepted for publication and has undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1111/hsc.12661

DR. MENG-WONG TAING (Orcid ID : 0000-0003-0686-2632)

Article type : Original Article

ABSTRACT We sought to examine the management and referral of non-healing mouth ulcer presentations in Australian community pharmacies in the Greater Brisbane region. Trained simulated patients visited 220 randomly selected community pharmacies within the Greater Brisbane region in 2016. Simulated patients enacted two non-healing (> 1 month) mouth ulcer scenarios: A direct product request (DPR) (n=110) and a symptom based request (SBR) (n=110). Results were documented and evaluated against Australian national pharmacy practice standards. Referral rates for pharmacy staff (pharmacist, pharmacy assistant or mixed – pharmacist and assistant) were also assessed. Australian pharmacy practice standards recommend staff ask six key questions during SBR and DPR consultations to enable informed decision making. Two questions relating to identifying the patient and their symptoms were asked in the majority of interactions (76% and 69% respectively); the remaining four questions relating to symptom duration, treatments tried, other medications and medical conditions were enquired in only 32%, 53%, 31% and 27% of interactions respectively. Simulated patients were referred to the doctor/dentist in only 11.8% of all interactions (both scenarios requiring referral). Overall staff handling of non-healing mouth ulcer consultations was suboptimal compared to national professional standards. In particular, duration of the non-healing mouth ulcer was enquired in less than one third of consultations potentially resulting in low referral rates by staff. This study identifies the need for increased oral cancer awareness and education for community pharmacy staff and reinforcing the importance of practising according to professional standards to effectively screen for potentially cancerous non-healing mouth lesions.

Keywords: Mouth neoplasms, referral consultation; service evaluation; community pharmacies

Author Manuscript

What is known about this topic: • Oral cacer is a coo aligacy orldide, rakig eighth ad thirteeth for ales ad feales respectiely. • No-healing ulcers persisting for longer than 2-3 weeks should be referred without delay to eliminate the possibility of a potentially cancerous lesion • Couity pharacy staff ca hae a iportat role i the early detectio ad preetio of oral cacer.

What this paper adds: • First Australia study iestigatig pharacy aageet of potetial outh cacers • Approximately 10% of staff referred a non-healing mouth ulcer; suboptimal handling of presentations compared to pharmacy professional standards • Need for icreased oral cacer aareess ad educatio for pharacy staff

INTRODUCTION In Australia, 2915 new cases of cancer in the oral cavity were diagnosed in 2014.(AIHW) Oral cancer is a common malignancy worldwide, ranking eighth and thirteenth for males and females respectively.(McCullough, Prasad, & Farah, 2010) (Farah, Simanovic, & Dost, 2014) Typically, patients diagnosed are over 40 years of age, and known causal risk-factors include alcohol, tobacco, human papillomavirus, micronutrient deficiency and betel quid use.(Farah, Vu, Allen, McCullough, & Ford, 2012; Ford & Farah, 2013) Importantly, early detection, diagnosis and treatment is known to significantly enhance survival rates and reduce morbidity.(Ford & Farah, 2013) Two reasons for delayed treatment include patient delay and professional delay.(Farah et al., 2012) Patient delay is defined as the period between recognition of signs/symptoms and first consultation with a health practitioner. Professional delay is

a consequence of delayed Author Manuscript referral to specialist medical or dental practitioners for definitive diagnosis of the tumour.(Gómez et al., 2010; Scott, Grunfeld, Main, & McGurk, 2006) Models have also been used to describe delays in detection and diagnosis with authors Emery et al. proposing a recently refined version of the Anderson model for total patient delay.(Walter, Webster, Scott, & Emery, 2012) This model describes four time intervals (appraisal, help- seeking, diagnostic and pre-treatment) making up the total time between the appearance of signs/symptoms of a

This article is protected by copyright. All rights reserved cancer and the commencement of treatment.(Ford & Farah, 2013) The study by Vaughan et al. reports the largest proportion of total delay time arises from patients not presenting to their dentist or doctor.(S. N. Rogers et al., 2007) Factors associated with patient delay include low socioeconomic status, heavy use of alcohol and tobacco, limited access to primary health care and self-medication.(Farah et al., 2012; van der Waal, de Bree, Brakenhoff, & Coebergh, 2011)

Community pharmacy staff are in a position to influence patient delay and can have an important role in the early detection and prevention of oral cancer. There are numerous types of oral cancer, with oral squamous cell carcinoma being the most common form. Early signs and symptoms of oral cancer prior to malignant change can include ulceration, changes in texture/colour, swelling, and precursor lesions (leukoplakia, erthroplakia). The most common oral symptoms noticed by patients are persistent lumps, soreness in the oral cavity, abscess and a non- healing sore/ulcer.(Farah et al., 2012) Patients commonly attribute these symptoms to a dental problem, infection or problems with a prosthesis and unfortunately, approximately half of oral cancers worldwide are diagnosed at advanced stages III and IV when the five year survival rate is about 50%, but can be as low as 15%.(van der Waal et al., 2011) More than 75% of UK pharmacists receive questions relating to oral enquiries on a regular basis, with the most frequent concerning mouth ulcers and toothaches.(Dickinson, Howlett, & Bulman, 1995) A study in Johannesburg reported similar findings with mouth ulcers the most common reason for seeking advice (88%).(Gilbert, 1998) A recent Australian study found that more than half of community pharmacists and assistants are involved in identifying signs and symptoms for oral health problems.(Freeman, Abdullah, Ford, & Taing, 2017; Taing, Ford, Gartner, & Freeman, 2016) People with non-healing ulcers may seek advice for management (symptom-based presentation), or directly request products for treatment (direct product request). Community pharmacy staff however, may not be appropriately assessing or managing non-healing mouth ulcers. A UK study reported only 10% of pharmacists referred a patient to a doctor or dentist presenting symptoms suggestive of oral cancer.(Scully, Gill, & Gill, 1989) A study in the US identified only one-third of pharmacists appropriately referred patients presenting with a non-healing ulcer to dentists/doctors, whereas two-thirds recommended purchase of an over the counter product.(Leonard, Isetts, & Leonard, 1996) More recently, studies in the UK and Spain reported higher rates of referral with up to 82% of pharmacist referring a non-healing mouth ulcer.(Evans & Gibbons, 2005; Varela-Centelles et al., 2012) A UK study has shown that pharmacist training for the detection of potentially malignant oral lesions effectively improves referral rates.(S. Rogers, Lowe, Catleugh, & Edwards, 2010) Currently, no studies have evaluated how Australian pharmacy staff manage non-healing mouth ulcer presentations, or whether they appropriately refer these patients. This study is exploratory and will describe pharmacy staff management of non-healing mouth ulcers in two Author Manuscript simulated scenarios; a patient symptom-based presentation (SBR) and a direct product request (DPR). This study will determine whether Australian community pharmacy staff potentially contribute towards patient diagnostic delay of non-healing mouth lesions using standardized patient methodology.

METHODS Simulated patient methodology Simulated patient methodology is commonly used to covertly observe pharmacy staff practices with staff unaware that an evaluation is taking place.(Watson, Norris, & Granas, 2006; Xu, de Almeida Neto, & Moles, 2012) Simulated patients consisted of trained individuals acting as consumers to observe interactions with pharmacy staff when presented with either a symptom based (SBR) or direct product request (DPR). Simulated patients presented to pharmacies with standardised requests (Table 1) and recorded details of the interaction immediately after leaving each pharmacy. This method is used to minimise the Hawthorne effect (participants changing their behaviour as a natural response from knowing they are being observed) whilst assessing pharmacy practices.(Byrne, Wood, & Spark, 2017; Kashyap, Nissen, Smith, & Kyle, 2014) Ethics approval for this study was granted by the Bellberry Human Research Ethics Committee in March 2016 (Application No: 2016-01-029).

Pharmacy selection The Greater Brisbane Region (GBR) is a geographical area defined by the Australian Bureau of Statistics as representing the functional extent of Brisbane City and is used for the collation of social and economic survey data(Statistics, 2012, 2016). This region extends beyond the city itself to include a number of rural areas and is home to approximately 2.3 million people. Publically available online directories (Yellow Pages and Google Maps) were utilised to create an excel spreadsheet database containing contact details of all community pharmacies located within the GBR (400 pharmacies). All community pharmacies within the database were sent an email notifying them of the study and informing them that they may be visited by a simulated patient within the following 6 months. The email provided pharmacies an opportunity to opt-out from the study via return email or fax; pharmacies that did not opt-out were automatically included in the study. Pharmacies were randomised (Microsoft Excel 2010) to produce a sample of 110 pharmacies for each scenario (220 pharmacies in total). The sample size of 220 out of a total of 400 pharmacies was determined based on a precision of 5% for population percentage estimates with a 95% level of confidence.(Krejcie & Morgan, 1970)

Data collection Community pharmacies were visited between March and May 2016. Two simulated female patients of similar age (21 and 22 years old) with reasonable emotional maturity, intelligence and trustworthiness were selected.(Watson et al., 2006) To maintain consistency during scenario replication, each simulated patient was trained to deliver one scenario (SBR or DPR) by authors (MWT and CF) through simulated mock-interaction sessions to present scenarios and collect data in a consistent manner. Simulated patients provided pharmacy staff with information (Table 1) only Author Manuscript if requested. Results from the interaction were recorded immediately after visiting each pharmacy using a standardised data collection form. The appropriate outcome for both scenarios was non-delayed/immediate referral to a dental or general practitioner as both scenarios described a non-healing mouth ulcer (with symptoms and risks factors for oral cancer). The data collection form recorded three separate categories for referrals; immediate/non- delayed referral (i.e. patient advised to see their doctor), non-urgent referrals defined as delayed referrals (e.g. see a This article is protected by copyright. All rights reserved doctor if the ulcer does not heal/use the product for five days and then see a doctor if the ulcer has not improved) and no referral. Prior to consultation if the simulated patient recognised any staff working within the pharmacy, the simulated patient left the premises to prevent potential biases during the interaction. The pharmacy was excluded and another pharmacy from the randomised list was substituted until 110 community pharmacies were visited by each simulated patient. After completion of the data collection form (outside the pharmacy), one staff member involved in the interaction was informed of the study and provided an educational fact sheet to assist best practice towards the management of mouth ulcer presentations in community pharmacy.

Data Analysis Data from standardised collection forms were entered into Checkbox Survey (v4.7) and exported into Statistical Package for Social Sciences (SPSS v24) for descriptive analysis. Chi-squared analyses were performed and Fisher’s exact test was applied if the expected number of observations per cell were fewer than five.(Bower, 2003; McDonald, 2009) Comparison of proportion estimates between staff and scenario types were made with the significance level set at P < 0.05. When contingency tables were larger than 2x2, adjusted standardised residuals was used to show cells which had larger or smaller counts than expected if two variables were considered independent, with a significance level of P < 0.05.(IBM Support, 2016) Multiple linear regression analysis was conducted to determine which explanatory variables (pharmacy type, pharmacy socio-economic indexes for areas (SEIFA), QCPP accreditation, approximate consultation length, staff type, gender and scenario) best predicted the number of questions asked by pharmacy staff during consultations. For the purpose of exploratory model building, the stepwise method for variable selection was performed.(Bland, 2015) Dummy variables were created for all categorical variables to allow meaningful comparisons with a reference category. The final model was tested for independence of residuals, collective linearity between dependent and independent variables, homoscedasticity, multicollinearity and outliers. Logistic regression was also performed to determine associations between demographic characteristics and whether pharmacy staff enquired about the duration of the mouth ulcer and referred without delay (best-practice)(eTGcomplete[Internet]. 2015; Sansom LN, 2015). A P value < 0.05 was considered statistically significant. In this study, staff interactions were classified as pharmacist interactions (involving only pharmacists), pharmacy assistant interactions (no pharmacist direction/pharmacist supervision) or mixed interactions (involved consultation with an assistant and pharmacist). Additionally, intern pharmacists were included in the pharmacist category for analyses (given their training is more closely aligned with pharmacists compared with assistants). The professional role of the staff member was identified either by uniform and/or presence of name badges.

RESULTS Author Manuscript A total of 36 pharmacies opted-out from the simulated patient study. No demographic correlations between pharmacies opting-out from the study compared to those that did not were observed. Common reasons for choosing to opt-out included: already involved in mystery shopping through Guild/franchise agreements and currently training new staff and it would not be suitable to assess them, 16 pharmacies did not provide a reason.

Pharmacy characteristics and staff involved during the interactions are shown in Table 2. The majority (80.9%) of interactions were completed within 1-5 minutes. A significantly greater number of QCCP accredited pharmacies were visited in DPR (92.7%) vs SBR (65.5%) presentations (p<0.0001). QCCP pharmacies were more likely to ask about symptoms present during SBR (p<0.029); however no other significant differences were found. Differences were observed between staff types for SBR and DPR scenarios/interactions; 20 pharmacy staff who could not be identified during SBR interactions were excluded in subsequent analyses comparing staff type (pharmacist, pharmacy assistant and mixed interactions) questioning and treatment recommendations. Australian pharmacy practice standards recommend pharmacy staff ask a minimum of 6 key questions during SBR and DPR consultations to enable informed decision making regarding appropriate treatment/advice.(S. Benrimoj & Wilson, 2006) Results from staff questioning/information gathering for standardised patient requests are shown in Table 3. Of the six key questions, three were asked in more than 50% of interactions (who is the patient, 76.4%; actual symptoms, 68.6% and other medications, 52.7%). Risk factors for oral cancer of patient age, smoking status and alcohol intake were asked in 6.8%, 3.6% and 0% of all interactions respectively. Duration of the lesion was determined in 27.3% of all interactions.

During mixed interactions (involving consultation from both a pharmacist and pharmacy assistant), staff were significantly more likely to ask five of the six key questions (Table 3). Adjusted residuals analysis showed pharmacists were no more or less likely to elicit information from the simulated patient compared to other staff types; however, pharmacy assistants were less likely to determine what the actual symptoms were, whether the patient had tried anything for the ulcer, whether the patient took any other medications and the duration of symptoms. No significant differences were observed between staff types relating to enquiring about risk factors for oral cancer. Comparing SBR and DPR interactions, pharmacy staff were significantly more likely to determine who the patient was in the DPR scenario, whereas investigation of symptoms, previously used treatments and age of the patient were asked more frequently in SBRs (Table 3).

Overall, 11.8% of all interactions were referred without delay to either a dental or medical practitioner, with the majority of staff (63.6%) not referring the patient (Table 4). No significant differences in treatment recommendations was observed between SBR and DPR scenarios. Adjusted residuals analysis however showed that pharmacy assistants were significantly less likely to refer without delay compared to other staff types and more likely not to refer the simulated patient. Mixed interactions were significantly more likely to provide non-urgent referrals compared to other staff types.

Author Manuscript Two variables were significantly associated with the number of questions asked during staff interactions. These variables were approximate consultation length i) 1-5 minutes (compared to reference category <1 minute) and ii) 5- 10 minutes (compared to reference category <1 minute). Interactions lasting between 1-5 minutes were likely to elicit an additional 2 questions, and interactions lasting between 5-10 minutes an additional 5 questions compared to consultations of less than 1 minute (the regression coefficient and standard error for each variable are shown in This article is protected by copyright. All rights reserved

Supplementary table 1). Logistic regression was used to determine associations between staff demographic characteristics and best practice (enquiry about the duration of the mouth ulcer and referral without delay). Associations were adjusted for pharmacy type, pharmacy socio-economic indexes for areas (SEIFA), QCPP accreditation, approximate consultation length, staff type, gender and scenario (Supplementary table 2). Consultation time was not a significant predictor for best practice. Pharmacists were significantly more likely to ask the duration of the ulcer and provide an immediate referral to a dental or medical practitioner (best practice) compared to pharmacy assistants.

DISCUSSION This is the first Australian study to investigate how community pharmacy staff manage a non-healing (and potentially cancerous) mouth ulcer presentation. The study simulated two non-healing mouth ulcer scenarios (SBR and DPR) to community pharmacies located within the Greater Brisbane region and assessed information gathering practices and treatment recommendations by staff which was compared to Australian professional practice standards and guidelines). This study is also the first study that the authors are aware of to assess mixed interactions (ie. an interaction involving consultation with an assistant and pharmacist direction/personal supervision) in comparison to pharmacist/pharmacy assistant-only interactions.

Australian pharmacy professional practice standards recommend pharmacy staff ask a minimum of six key questions during SBR and DPR consultations to enable appropriate and informed decision making regarding advice/treatment.(S. Benrimoj & Wilson, 2006) Findings from this study indicate that mixed interactions resulted in significantly higher enquiries for five out of six key questions compared to pharmacist or pharmacy assistant-only interactions. It is unclear why during consultations, mixed interactions with the simulated patient elicited more information compared to pharmacist and assistant-only interactions however, one explanation may be that multidisciplinary/multi-professional teams are more conducive towards fostering comprehensive and improved patient enquiry approaches and processes. Carrier & Kendall describe inter-discipliary ork as a illigess to share and indeed give up exclusive claims to specialist knowledge and authority, if the needs of clients can be met ore effectiely y other professioal groups.(Carrier & Kendall, 1995) Community pharmacies are mostly comprised of small inter-professional teams including pharmacists and pharmacy assistants that have complementary skills/roles who are committed to patient-centred care and hold themselves mutually accountable. Some advantages of effective inter-professional teams discussed in literature include enhancing information sharing, delivering services that are planned and co-ordinated, streamlining work practices and acquisition of new skills

leading to increased job satisfaction.(HornbyAuthor Manuscript & Atkins, 2008; May, 1997; Onyett & Smith, 1998) These theories may partly explain why mixed interactions in this study led to improved patient assessment processes compared to pharmacist/assistant-only interactions. There is however debate whether better patient assessment processes leads to improvements in clinical outcomes.(Pillay et al., 2016)

The results from this study show that staff questioning processes (irrespective of the scenario presented) for non- healing mouth ulcer presentations was sub-optimal when compared to practice standards. Recent Australian simulated patient-based studies evaluating a range of presentations in community pharmacies (such as common ailments, insomnia and analgesics containing codeine) highlights approximately half or less of participants practice according to standards, with pharmacy assistants least likely to adhere.(Byrne et al., 2017; Collins et al., 2017; Kashyap et al., 2014) The integration of prescribed standards in practice is known to differ between community pharmacies. Approximately 90% of Australian community pharmacies are Quality Care Pharmacy Program (QCCP) accredited, which aims to ensure application of standards and professional advice is appropriate and consistent.(The Pharmacy Guild of Australia) Previous pharmacy simulated-patient studies did not report/differentiate between accredited vs non-accredited pharmacies, hence we are unable to discern whether this variable influenced staff adherence with national standards. This study is the first to demonstrate suboptimal adherence towards standards relating to non-healing mouth ulcer presentations regardless of the accreditation status of community pharmacies (located in the Greater Brisbane region).

It is estimated that Australian pharmacies perform almost 500,000 interventions yearly when dealing with OTC medicines, with one in five being clinically significant interventions (potentially life-saving/averting harm/emergency attention).(S. I. Benrimoj et al., 2005) Community pharmacy staff have an important role when dealing with nonprescription presentations and medicines especially from a quality assurance and risk management perspective. (Byrne et al., 2017; Kashyap et al., 2014; Singleton, 2013; Stupans, Owen, Ryan, Woulfe, & McKauge, 2010) In this study, pharmacy staff were presented with two scenarios (SBR and DPR) involving a solitary persistent and non- healing (4 weeks) lingual ulcer in a middle/older-aged male. This is probably the most frequent presentation of oral cancer, particularly given the patient’s history of risk factors including heavy tobacco smoking and alcohol consumption.(Farah et al., 2012; Mashberg & Samit, 1995; Scully et al., 1989) Appropriate management of non- healing ulcer presentations for non-dental health professionals requires determining the duration of the non-healing ulcer, and those persisting for longer than 2-3 weeks being referred without delay for further investigation to eliminate the possibility of a potentially cancerous lesion.(eTGcomplete[Internet]. 2015; Scott, Grunfeld, Auyeung, & McGurk, 2009; Scully et al., 1989) In this study, pharmacy staff asked about the duration of the ulcer in only 27% of interactions, potentially resulting in only 12% of all interactions being referred to a dental/medical practitioner. Logistic regression showed pharmacists were significantly more likely to enquire about the duration of the ulcer and refer the patient compared to pharmacy assistants (17% vs 4% respectively), however these referral rates are unacceptably low. Overall, results from this study indicate that the majority of pharmacy staff are not managing non- healing mouth ulcer presentations in community settings appropriately. This is in contrast to literature describing a Author Manuscript very high proportion of pharmacy staff (>95%) reporting confidence in identifying and managing mouth ulcer presentations.(Freeman et al., 2017; Taing et al., 2016) Given these conflicting findings, future studies utilising methodologies including standardised patient and/or case vignettes are necessary to determine whether Australian pharmacy staff are appropriately identifying and managing other common conditions.

Early detection, diagnosis and treatment of oral cancer is known to significantly enhance survival rates and reduce morbidity.(Ford & Farah, 2013) Pharmacists may be the first healthcare professional to assess and counsel patients for non-healing mouth ulcer presentations given patients regularly visit pharmacies with enquiries relating to ulcers.(Freeman et al., 2017; Taing et al., 2016) The findings from this study indicate pharmacy staff need further education about how to manage mouth ulcer conditions, with increased emphasis on clinical features. There is also a need for increased oral cancer awareness in addition to re-enforcing the importance of practising according to professional standards to effectively screen for suspicious and potentially cancerous lesions. Oral cancer awareness campaigns and other appropriate professional development targeting pharmacy staff may be useful interventions. Enhancing collaborations between pharmacy, medical and dental professional organisations may lead to better integrated and improved service delivery models.

A valuable aspect of this study is that it is the first to compare staff practices between pharmacists, pharmacy assistants and mixed-interactions, which is reflective of professional team-based interactions that commonly takes place during patient consultations. Another strength of this study is its large sample size. The covert nature of simulated patient methodology employed in this study also reduced biases resulting from the Hawthorne effect (in this study no simulated patients reported also being detected) and findings represent current pharmacy practices. A limitation of this study is that differences were observed between staff types involved in SBR and DPR interactions, which may influence interpretation of findings when comparing SBR and DPR scenarios. Findings from this study cannot be generalised to all community pharmacies given the small geographical region in which the study was conducted. This study also depended on the memory/recall of simulated patients which may have affected the accuracy of data collected. This was mitigated through comprehensive training support for simulated patients to present scenarios and collect data in a consistent manner; and immediate documentation of results following each interaction. In this study, it was not possible to identify the role of staff in all the interactions, as identification was determined by name badges and/or uniforms. This is a universal limitation when utilising simulated-patient methodologies.(Kelly, Williams, & Benrimoj, 2009; Norris, 2004; Schneider, Everett, Geelhoed, Kendall, & Clifford, 2009) Interactions where staff roles could not be identified were subsequently excluded from analyses comparing staff type questioning and treatment recommendations.

CONCLUSION Community pharmacy staff handling of non-healing mouth ulcer consultations was suboptimal compared to national professional standards. In particular, the duration of non-healing mouth ulcer was enquired in less than one third of consultations potentially resulting in low referral rates by staff. This may contribute to patient diagnostic delay of Author Manuscript cancerous oral lesions. There is a need for increased oral cancer awareness and education for community pharmacy staff in addition to reinforcing the importance of practising according to professional standards to effectively screen for cancerous mouth ulcers/lesions.

REFERENCES AIHW. Cancer in Austrlalia an overview 2014 Australian Institute of Health and Welfare’s Cancer. Canberra. Benrimoj, S., & Wilson, F. (2006). Standards for the provision of pharmacy medicines and pharmacist only medicines in community pharmacy. Revised, November 2005. Professional practice standards (Version 3). Canberra: Pharmaceutical Society of Australia. Benrimoj, S. I., Gilbert, A., Emmerton, L., Williams, K., Gadiel, D., Taylor, R., & Carroll, P. (2005). A cost-benefit analysis of Pharmacist Only (S3) and Pharmacy (S2) Medicines and risk-based evaluation of the standards. Bland, M. (2015). An introduction to medical statistics: Oxford University Press (UK). Boer, K. M. 00. Whe to use Fisher’s eact test. American Society for Quality Six Sigma Forum Magazine, 2, p. 35-37. Byrne, G. A., Wood, P. J., & Spark, M. J. (2017). Non-prescription supply of combination analgesics containing codeine in community pharmacy: A simulated patient study. Research in Social and Administrative Pharmacy. Carrier, J., & Kendall, I. (1995). Professionalism and interprofessionalism in health and community care: some theoretical issues Interprofessional issues in community and primary health care (pp. 9-36): Springer. Collins, J. C., Schneider, C. R., Faraj, R., Wilson, F., de Almeida Neto, A. C., & Moles, R. J. (2017). Management of common ailments requiring referral in the pharmacy: a mystery shopping intervention study. International Journal of Clinical Pharmacy, 39(4), 697-703. doi: 10.1007/s11096-017-0505-8 Dickinson, C., Howlett, J. A., & Bulman, J. S. (1995). The role of the community pharmacist as a dental health adviser. Community Dent Health, 12(4), 235-237. eTGcomplete[Internet]. (2015). Oral and Dental. from Therapeutic Guidelines Limited Evans, M., & Gibbons, A. (2005). Advice given in community pharmacies to patients with possible oral carcinoma. British Journal of Oral and Maxillofacial Surgery, 43(3), 253-255. Farah, C., Simanovic, B., & Dost, F. (2014). Oral cancer in Australia 1982–2008: a growing need for opportunistic screening and prevention. Australian Dental Journal, 59(3), 349-359. Farah, C., Vu, A., Allen, K., McCullough, M., & Ford, P. (2012). Oral cancer and potentially cancerous lesions-early detection and diagnosis: INTECH Open Access Publisher. Ford, P. J., & Farah, C. S. (2013).Author Manuscript Early detection and diagnosis of oral cancer: Strategies for improvement. Journal of Cancer Policy, 1(1–2), e2-e7. doi: http://dx.doi.org/10.1016/j.jcpo.2013.04.002 Freeman, C. R., Abdullah, N., Ford, P. J., & Taing, M.-W. (2017). A national survey exploring oral healthcare service provision across Australian community pharmacies. BMJ Open, 7(9). doi: 10.1136/bmjopen-2017-017940

Gilbert, L. (1998). The role of the community pharmacist as an oral health adviser--an exploratory study of community pharmacists in Johannesburg, South Africa. SADJ : journal of the South African Dental Association = tydskrif van die Suid-Afrikaanse Tandheelkundige Vereniging, 53(8), 439-443. Gómez, I., Warnakulasuriya, S., Varela-Centelles, P. I., López-Jornet, P., Suárez-Cunqueiro, M., Diz-Dios, P., & Seoane, J. (2010). Is early diagnosis of oral cancer a feasible objective? Who is to blame for diagnostic delay? Oral Diseases, 16(4), 333-342. doi: 10.1111/j.1601-0825.2009.01642.x Hornby, S., & Atkins, J. (2008). Collaborative care: interprofessional, interagency and interpersonal: John Wiley & Sons. IBM Support. (2016). Interpreting adjusted residuals in Crosstabs cell statistics. from https://www- 304.ibm.com/support/docview.wss?uid=swg21479605 Kashyap, K. C., Nissen, L. M., Smith, S. S., & Kyle, G. (2014). Management of over-the-counter insomnia complaints in Australian community pharmacies: a standardized patient study. Int J Pharm Pract, 22(2), 125-134. doi: 10.1111/ijpp.12052 Kelly, F. S., Williams, K. A., & Benrimoj, S. I. (2009). Does advice from pharmacy staff vary according to the nonprescription medicine requested? Ann Pharmacother, 43(11), 1877-1886. doi: 10.1345/aph.1L121 Krejcie, R. V., & Morgan, D. W. (1970). Determining sample size for research activities. Educ Psychol Meas. Leonard, M. S., Isetts, B. J., & Leonard, C. X. (1996). Response of community pharmacists to potential oral carcinoma. Journal of the American Pharmaceutical Association (Washington, DC: 1996)(3), 203-205. Mashberg, A., & Samit, A. (1995). Early diagnosis of asymptomatic oral and oropharyngeal squamous cancers. CA Cancer J Clin, 45(6), 328-351. May, C. (1997). Planning and managing teams. Health & Social Care in the Community, 5(4), 269-283. doi: 10.1111/j.1365-2524.1997.tb00123.x McCullough, M. J., Prasad, G., & Farah, C. S. (2010). Oral mucosal malignancy and potentially malignant lesions: an update on the epidemiology, risk factors, diagnosis and management. Aust Dent J, 55 Suppl 1, 61-65. doi: 10.1111/j.1834-7819.2010.01200.x McDonald, J. H. (2009). Handbook of biological statistics (Vol. 2): Sparky House Publishing Baltimore, MD. Norris, P. (2004). Reasons why mystery shopping is a useful and justifiable research method. Pharmaceutical journal, 272(7303), 746-747. Onyett, S., & Smith, H. (1998). The structure and organisation of community mental health teams. Serious Mental Health Problems in the Community. Policy, Practice and Research, 62-86. Pillay, B., Wootten, A. C., Crowe, H., Corcoran, N., Tran, B., Bowden, P., . . . Costello, A. J. (2016). The impact of multidisciplinary team meetings on patient assessment, management and outcomes in oncology settings: A Author Manuscript systematic review of the literature. Cancer Treatment Reviews, 42, 56-72. doi: http://dx.doi.org/10.1016/j.ctrv.2015.11.007 Rogers, S., Lowe, D., Catleugh, M., & Edwards, D. (2010). An oral cancer awareness intervention in community pharmacy. British Journal of Oral and Maxillofacial Surgery, 48(7), 498-502.

Rogers, S. N., Pabla, R., McSorley, A., Lowe, D., Brown, J. S., & Vaughan, E. D. (2007). An assessment of deprivation as a factor in the delays in presentation, diagnosis and treatment in patients with oral and oropharyngeal squamous cell carcinoma. Oral Oncol, 43(7), 648-655. doi: http://dx.doi.org/10.1016/j.oraloncology.2006.08.001 Sansom LN. (2015). Australian pharmaceutical formulary and handbook (23rd ed.). Canberra: Pharmaceutical Society of Australia. Schneider, C. R., Everett, A. W., Geelhoed, E., Kendall, P. A., & Clifford, R. M. (2009). Measuring the assessment and counseling provided with the supply of nonprescription asthma reliever medication: a simulated patient study. Ann Pharmacother, 43(9), 1512-1518. doi: 10.1345/aph.1M086 Scott, S. E., Grunfeld, E. A., Auyeung, V., & McGurk, M. (2009). Barriers and triggers to seeking help for potentially malignant oral symptoms: implications for interventions. J Public Health Dent, 69(1), 34-40. doi: 10.1111/j.1752-7325.2008.00095.x Scott, S. E., Grunfeld, E. A., Main, J., & McGurk, M. (2006). Patient delay in oral cancer: a qualitative study of patients' experiences. Psychooncology, 15(6), 474-485. doi: 10.1002/pon.976 Scully, C., Gill, Y., & Gill, Z. (1989). How community pharmacy staff manage a patient with possible oral cancer. British Journal of Oral and Maxillofacial Surgery, 27(1), 16-21. Singleton, J. (2013). Feature-Business models: Differentiating from the differentiators. AJP: The Australian Journal of Pharmacy, 94(1115), 62. Statistics, A. B. o. (2012). Statistical Geography. Canberra. Statistics, A. B. o. (2016, 23 September 2016). Greater Brisbane (GCCSA). Retrieved 20 October, 2016 Stupans, I., Owen, S., Ryan, G., Woulfe, J., & McKauge, L. (2010). Scaffolding patient counselling skills in Australian university pharmacy programs. Asia-Pacific Journal of Cooperative Education, 11(2), 29-37. Taing, M.-W., Ford, P. J., Gartner, C. E., & Freeman, C. R. (2016). Describing the role of Australian community pharmacists in oral healthcare. International Journal of Pharmacy Practice, n/a-n/a. doi: 10.1111/ijpp.12241 The Pharmacy Guild of Australia. Quality Care Pharmacy Program. 2017, from http://www.qcpp.com/about- qcpp/what-is-qcpp van der Waal, I., de Bree, R., Brakenhoff, R., & Coebergh, J.-W. (2011). Early diagnosis in primary oral cancer: is it possible? Medicina Oral Patología Oral y Cirurgía Bucal, 16(3), e300-e305. Varela-Centelles, P., Pedrosa, R., Lopez-Nino, J., Sanchez, M., Gonzalez-Mosquera, A., Mendez, A., & Seoane, J. (2012). Oral cancer awareness at chemist's and herbalist's shops: new targets for educational interventions to prevent diagnostic delay. Oral Oncol, 48(12), 1272-1275. doi: 10.1016/j.oraloncology.2012.06.008 Walter, F., Webster, A., Scott, S., & Emery, J. (2012). The Andersen Model of Total Patient Delay: a systematic review Author Manuscript of its application in cancer diagnosis. Journal of Health Services Research & Policy, 17(2), 110-118. doi: 10.1258/jhsrp.2011.010113 Watson, M., Norris, P., & Granas, A. (2006). A systematic review of the use of simulated patients and pharmacy practice research. International Journal of Pharmacy Practice, 14(2), 83-93.

Xu, T., de Almeida Neto, A. C., & Moles, R. J. 0. A systeatic reie of siulated‐patiet ethods used i couity pharacy to assess the proisio of o‐prescriptio edicies. International Journal of Pharmacy Practice, 20(5), 307-319.

Table1: Scenarios Table 2: Pharmacy and pharmacy staff demographics Table 3: Staff questioning Table 4: Treatment Recommendations made by pharmacy staff Supplementary Table 1: Factors that influence the number of questions asked by pharmacy staff during consultations from multiple linear regression analysis (n=200) Supplementary table 2: Associations between demographic characteristics and staff enquiring about the duration of the mouth ulcer and referred without delay Author Manuscript

Table1: Scenarios University student in her 20s:

SBR: Can you please provide advice about DPR: Can I please buy Kenalog (Kenalog in managing mouth ulcers? Orabase) for a mouth ulcer?

If asked:

- The advice/product is for their father - Father is 60 years old - Mouth ulcer information: o located on his tongue o has been there for a month (4 weeks) o not originally painful but is now sore o unsure what ulcer looks like o cause of ulcer is unknown, no known trauma (e.g. biting tongue) before it started o has not seen a Doctor or Dentist about the mouth ulcer o has not tried any treatments for the ulcer o Father does not regularly experience mouth ulcers - Other medications used by their father: o occasioal Paadol paracetaol 500g for headaches - Other information: o Father has no allergies/bad reactions to a medication o father likes a drik, if asked specifically, he is a ‘pretty heavy drinker’ (> 4 drinks per day) o father is a heavy smoker, if asked specifically, smokes 1 pack (20 cigarettes) a day DPR Specifically: o The simulated patient (daughter) is recommending Kenalog for her dad (she used it for a mouth ulcer a year ago) and it helped o Father has never used Kenalog before Bolded information are symptoms and risk factors for oral cancer4,33 Author Manuscript

Symptom Based Direct Product P-value Total; n=220 Request; n=110 Request; n=110 n (%) n (%) n (%) Pharmacy Type

Independent 39 (35.5) 35 (31.8) 74 (33.6) 0.669 Chain/Banner 71 (64.5) 75 (61.2) 146 (66.4) Pharmacy Location

Shopping Centre 57 (51.8) 35 (31.8) 92 (41.8) Strip 45 (40.1) 54 (49.1) ND 99 (45.0) Medical Centre 42 (38.2) 20 (18.2) 62 (28.2) Stand alone 1 (0.9) 8 (7.3) 9 (4.1) Pharmacy SEIFA

Low (1-4) 38 (34.5) 31 (28.2) 69 (31.4) Medium (5-7) 24 (21.8) 21 (19.1) 0.401 45 (20.5) High (8-10) 48 (43.6) 58 (52.7) 106 (48.2) QCCP accredited

Yes 72 (65.5)* 102 (92.7)* 174 (79.1) <0.001 No 38 (34.5)* 8 (7.3)* 46 (20.9) Approximate Consultation Time (mins) < 1 5 (4.5) 9 (8.2) 14 (6.4) 1 – 5 92 (83.6) 86 (78.2) 178 (80.9) 0.515 6 – 10 13 (11.8) 15 (13.6) 28 (12.7) > 11 0 (0.0) 0 (0.0) 0 (0.0) Staff Gender

Female 84 (76.4) 89 (80.9) 0.511 173 (78.6) Staff Type

Mixed Interaction˄ 32 (29.1) 20 (18.2) 52 (23.6) Pharmacist 25 (22.7)* 41 (37.3)* 66 (30.0)

Intern Pharmacist Author Manuscript 2 (1.8) 1 (0.9) <0.001 3 (1.4) Pharmacy assistant/technician 31 (28.2)* 48 (43.6)* 79 (35.9) Other/unsure 20 (18.2)* 0 (0.0)* 20 (9.1)

Notes:

Pharmacy location variable column total adds to > 100%, which indicates some pharmacies where a composite of several categories. Gender variable refers to the staff member providing the final treatment recommendation. ˄ A mixed interaction was defined as an interaction involving consultation with an assistant and pharmacist direction/personal supervision. ND – Probability not determined as pharmacy location categories were not mutually exclusive. * Cells having larger or smaller counts/proportions than expected if two variables were considered independent, , with a significance level of p<0.05 Table 2: Pharmacy and pharmacy staff demographics Author Manuscript

Table 3: Staff questioning

Staff Type Scenario Total Questions (n=200) (n=220) (n=220) Pharmacy Pharmacist Mixed SBR DPR assistant P P (n =69) (n=79) (n=52) value (n=110) (n=110) value

n (%) n (%) n (%) n (%) n (%) n (%) Who is the 60 (87.0) 59 (74.7) 39 (75.0) 0.150 67 (60.9) * 101 (91.8) * <0.001 168 (76.4) patient?˄

Actual symptoms- 52 (75.4) 45 (57.0)* 44 (84.6)* 0.002 86 (78.2) * 65 (59.1) * 0.003 151 (68.6) what are they?˄

Tried anything for this mouth 19 (27.5) 18 (22.8)* 26 (50.0)* 0.003 58 (52.7) * 13 (11.8) * 0.000 71 (32.3) ulcer?˄

Other 39 (56.5) 30 (38.0)* 35 (67.3)* 0.003 56 (50.9) 60 (54.5) 0.685 116 (52.7) medications?˄

Medical 20 (29.0) 19 (24.1) 24 (46.3)* 0.024 28 (25.5) 39 (35.5) 0.143 67 (30.5) conditions?˄

Duration of 16 (23.2) 12 (15.2)* 25 (48.1)* 0.0001 35 (31.8) 25 (22.7) 0.173 60 (27.3) symptoms?˄

Age of 4 (5.8) 2 (2.5) 3 (5.8) 0.521 13 (11.8) * 2(1.8) * 0.006 15 (6.8) Patient?#

Smoking 3 (4.3) 1 (1.3) 4 (7.7) 0.156 4 (3.6) 4 (3.6) 1 8 (3.6) status?#

Alcohol Author Manuscript 0 (0.0) 0 (0.0) 0 (0.0) ND 0 (0.0) 0 (0.0) ND 0 (0.0) intake?#

˄ 6 key questions: Australian standards for the provision of pharmacy medicines and pharmacist only medicines in community pharmacy34

* Cells having larger or smaller counts/proportions than expected if two variables were considered independent, with a significance level of p<0.05 # Risk factors for oral cancer.4,35 ND – Not determined

Author Manuscript

Table 4: Treatment Recommendations made by pharmacy staff

Treatment Staff Type Scenario Total recommendations (n=200) (n=220) (n=220)

Pharmacy Pharmacist Mixed SBR DPR assistant P- P- value (n=69) (n=79) (n=52) value (n=110) (n=110) n (%) n (%) n (%) n (%) n (%) OTC medicines 67 (97.1) 76 (96.2) 47 (90.4) 0.245 103 (93.6) 105 (95.5) 0.768 208 (94.5) (S2/S3) Complementary 1 (1.4) 1 (1.3) 1 (1.9) 1 3 (2.7) 0 (0.0) 0.247 3 (1.4) medicines (CAMs) No referral ˃ 45 (65.2) 61 (77.2)* 24 (46.2)* 0.001 63 (57.3) 77 (70.0) 0.068 140 (63.6) Referral without 12 (17.4) 3 (3.8)* 8 (15.4) 0.021 16 (14.5) 10 (9.1) 0.296 26 (11.8) delay ˃ Non-urgent 12 (17.4) 15 (19.0) 20 (38.5)* 0.012 31 (28.2) 23 (20.9) 0.273 54 (24.5) referral ˃

* Cells having larger or smaller counts/proportions than expected if two variables were considered independent, with a significance level of p<0.05 ˃ Referral to a Dental or General Practitioner

Author Manuscript