Participant and GP perspectives and experiences of screening for undiagnosed type 2 diabetes in community pharmacy during the Pharmacy Diabetes Screening Trial

In Australia, an estimated 500,000 adults have undiagnosed type 2 diabetes mellitus (T2DM) and a further 2 million are predicted to have prediabetes [1‐3]. As the risk of complications for those with established diabetes can be reduced by enabling timely intervention through early case detection, screening is recommended in many national [4] and international guidelines [5]. Most screening for T2DM is delivered via general practitioners (GPs); however, approximately 15% of Australians did not visit a GP at all during 2018 (3.75 million). Social disadvantage, ethnicity, and remoteness are some of the reasons that people typically see their GP less often [6]. Low levels of GP contact are associated with sub-optimal health outcomes for patients [7] as well as providing less opportunity for screening service provision. Furthermore, even those who visit GPs regularly may experience suboptimal rates of T2DM and cardiovascular disease (CVD) risk factor detection and management, including vulnerable groups such as those with mental illness and kidney disease [8‐10].

In recent decades, increasing attention has focused on the need and opportunity to enhance access to prevention in primary care through integrated care with community pharmacy [11]. Community pharmacies are freely accessible, widely available and in frequent contact with individuals who are either healthy or unwell. This provides an opportunity to engage people along the health spectrum, including hard-to-reach populations who do not utilise other health services. Australians visit a community pharmacy an average of 14 times per year, the highest usage of all the major health care providers [12]. Moreover, evidence from a systematic review suggests that community pharmacies are feasible sites for screening and that a significant number of risk factors, such as high blood pressure, cholesterol, and diabetes risk, were correctly identified in community pharmacies [13].

Pharmacy-based screening services, therefore, can enhance access to screening as well as provide a gateway to GPs and other health services using a risk assessment and POC testing in a coordinated referral screening process. This aligns directly with the Australian National Diabetes Strategy [14] that aims to coordinate existing limited health care resources across all levels of government to reduce the impact of T2DM in the community. The recently completed Pharmacy Diabetes Screening Trial (PDST) was a landmark pharmacy trial, involving 14,089 participants at 339 sites across Australia that accounted for remoteness and socio-economic conditions. It compared the effectiveness and cost-effectiveness of three feasible approaches to screening for undiagnosed T2DM or risk of T2DM. Community pharmacies were selected from geographical groups of co-located postcodes and randomly allocated to three investigation groups, viz: (1) paper-based risk assessment alone using the Australian Diabetes Risk (AUSDRISK) score [15] (Group A); (2) AUSDRISK followed by a POC HbA1c test (Group B) if AUSDRISK score was elevated; and (3) AUSDRISK followed by a POC small capillary blood glucose test (scBGT) (Group C) if AUSDRISK score was elevated [16].

The protocol for referral of screening participants to their GP was adapted from the Royal Australian College of General Practitioners (RACGP) guidelines [4] through consultation with the PDST Expert Panel, which included national diabetes experts and key stakeholder representatives. In Group A, screening participants with an AUSDRISK score ≥ 12 were referred to the GP. In Group B, screening participants with an AUSDRISK ≥ 12 also received an HbA1c POC test in the pharmacy and those with an HbA1c ≥ 39 mmol/mol (5.7%) were referred to the GP. In Group C, screening participants with an AUSDRISK ≥ 12 also received a small capillary blood glucose test (scBGT) using a POC device in the pharmacy and those with either a fasting blood glucose (FBG) of ≥ 5.5 mmol/l or a random blood glucose (RBG) of ≥ 7.0 mmol/l were referred to the GP.

During the PDST, there were 145 confirmed newly diagnosed cases of T2DM and 338 cases of newly identified prediabetes. Consistent with one of the study hypotheses, of the three approaches to screening, the risk assessment using the AUSDRISK tool followed by a POC HbA1c test for those with AUSDRISK scores of ≥ 12, showed the highest overall detection rate of T2DM (1.5% of the total screened population) compared to Groups A (0.8%) and C (0.6%) [17].

The communication and support strategies used by the PDST pharmacies were developed in conjunction with the Expert Panel to facilitate acceptance of the diabetes screening service protocol by pharmacy clients and general practice. These included the provision of standardised letters to GPs about the trial and advice to all participating pharmacies to contact and disseminate this information to their local GP practices before commencing screening recruitment. Pharmacists delivering screening were supported by bespoke screening software (PDST Guildcare™) to ensure protocol-based screening, communication and referral, and a highly standardised approach across trial arms. A proforma referral letter that included the results of the screening was generated through the PDST Guildcare™ software for all screening participants requiring referral. Pharmacists were mandated to give a copy to each individual and to supply a copy to their nominated GP practice with their consent.

Ultimately, the effectiveness of such programs depends on the level of uptake of screening and referral by consumers and for those found to be at risk, the provision of diagnostic testing, and continuity of care by GPs. The aim of this study was therefore to explore consumer and GP experiences of the screening program and to identify factors influencing decisions to act on referrals. The level of referral uptake by screening participants together with the level of response from the GP served as a measure of how well the PDST screening service fitted into primary care.

This analysis was a secondary mixed methods study, embedded within the PDST involving GPs and screened participants in the PDST.

Between 1^st April and 21^st November 2017, 2,242 surveys were sent (via email, SMS or post) to referred participants and 369 responses were received (response rate 16%). During the same period, 2,989 surveys were emailed to non-referred participants and 520 responses were received (response rate 17%). Overall, survey responses were similar between the referred and non-referred participants and demonstrated an even spread of responses from the three investigation groups, by gender, by age, and by highest screening service component provided (i.e., risk assessment plus a POC test plus referral). Table 1 describes the demographics of the surveyed sample and compares it to the full trial population. The demographic profile of the survey sample closely matched that of the full trial population.

Based on screening participants’ self-reports obtained at the six-week and three-month follow-ups, an average of one in two had followed up with their GP; with uptake significantly higher for Group C (65%) than for Group B (56%) or Group A (44%) (p < 0.001).

In this paper we presented the experiences and perspectives of consumers and GPs in their engagement with the pharmacy diabetes screening service implemented in the PDST, a unique clustered RCT involving a nationally representative sample of Australian pharmacies and consumers aged 35–74 years. As key stakeholders, screening participant perspectives on the appeal, value and functioning of the pharmacy screening service is key to any future national rollout. Equally critical is the issue of providing continuity of care through successful integration of a pharmacy-based diabetes screening service into primary care.

Responses to the consumer survey indicated that the screening service was well received in the pharmacy by respondents, and this was independent of their trial arm or whether they were referred to the GP. Most survey respondents felt the pharmacist explained the results well, conducted the service in a professional manner and would support the future provision of such a screening service in community pharmacy. This concurs with findings of other studies of pharmacy screening services [20‐22].

While there was general satisfaction with the quality of the service, comments from the consumer surveys suggest that screening services without a finger-prick test were less valued by both consumers and pharmacists and might lead to less confidence in the result. If our findings are representative of the general view, it may help to explain increased numbers of pharmacy withdrawals and greater difficulty with participant recruitment in Group A (particularly for recruitment in rural and regional areas), where no option of finger prick testing existed. In fact, several pharmacies reported that some people demanded a POC test regardless of their AUSDRISK score and in these circumstances the pharmacists did not include them in the trial; this may have disproportionately affected Group A pharmacies. This corroborates with pharmacist feedback from previous screening trials, which showed that it may be difficult for some pharmacists to promote a paper-based service if consumers perceive it to be less comprehensive than what this pharmacy already offered [22, 23].

Referral uptake as an indirect measure of participant engagement and acceptance of the screening program also support the inclusion of a POC test as the preferred option. The provision of a referral by the pharmacist encouraged screened participants to visit their GP for follow-up thereby fostering continuity of care [24].

Another important component of the pharmacy diabetes screening service was provision of an approved information package and verbal counselling to encourage risk factor reduction by screened individuals at elevated risk. As shown in the results, a significant proportion of survey respondents had implemented a positive change in lifestyle (most likely to be increase in physical activity) to reduce their risk of T2DM. This concords with other findings in the literature demonstrating the capacity of community pharmacist interventions to encourage and realise healthy lifestyle behaviour changes in their clients [25‐27].

The evaluation of GP perspectives revealed that, at least among those GPs who responded to the survey, the PDST referral from the pharmacy worked well to inform them about their referred patient’s involvement in the trial and that the recommendations for follow-up care made by the pharmacists were appropriate. They highlighted the need for such a service, especially for people who do not visit their GP regularly but raised the issue of fragmentation/duplication of care for those who were frequent attendees at their practice, as has previously been reported in the literature [28]. To avoid duplication the pharmacist needs to know what tests have already been performed for an individual client. This may be addressed in the future by sharing of information via shared electronic records [29]. Moreover, patients need to be informed of their test status by GPs. During the PDST participants were asked before enrolment if they had been tested in the past 12 months, yet many were unaware of whether they had been tested and of the results of those tests. Equally, effective communication between pharmacists and GPs will ensure that prevention messages may be reinforced for patients who are at high risk of diabetes but are unaware of it. To facilitate continuity of care and integrate the pharmacy into the primary care network, a secure effective communication channel between GPs and pharmacists must also be developed, potentially via more effective functionality and widespread adoption of shared medical records such as the MyHealth record system.

As with most studies there were limitations. Whilst every effort was made to provide standardised approaches and clear documentation, as well as appropriate answer fields in the surveys to minimise errors in responses, it is acknowledged that some data obtained from consumers and GPs was incomplete. Furthermore, as the response rates were low at 15–17%, these perspectives may not be representative of the larger cohorts of non-responders.Changes in health behaviours were also based on self-report to the pharmacist during screening or reported anonymously by the respondent in the follow-up survey, which may have been subject to bias. Moreover, it is likely that non-responders of the referred patient sample were not as engaged and therefore less likely to go to their GP or think positively about the screening as part of the follow-up survey was about whether they went to their GP for follow-up (and whether they had changed their diet and exercise). One might also infer from the low GP response rate that there was limited awareness of the screening program among many GPs to whom referrals were addressed underscoring the ineffectiveness of communication between pharmacists and GPs. However, challenges with recruiting GPs to complete surveys are well known, and lower response rates than ours for issues that GPs would certainly be aware of are commonplace; hence we should not assume the low response rate is exclusively the result of lack of awareness [30, 31]. Perceptions that screening is beyond the scope of practice of pharmacists reported by a proportion of GP respondents may also have factored into the low response rate.

Tailored strategies to improve response rate to mail surveys may be appropriate to ensure external validity of findings for future surveys of this nature, and the evidence points to the utility of various strategies. For example, inclusion of monetary incentives has been shown to increase response rates [32]. Lotteries have also been examined as a possible incentive with mixed results. Provision of scratch lotteries was found to be the most effective of this form of incentivisation [33]. A study which compared follow-up mailings and monetary incentives to maximize response rates found both to be effective. However, in the context of budgetary restraints follow-up mailings are preferred over monetary incentives. If there is limited time for survey administration, monetary incentives may be preferred over follow-up mailings [34]. In our study due to resource constraints, we were not able to deploy these types of incentives.

As the prevalence of T2DM continues to rise, there is a need for novel programs that provide effective screening for undiagnosed T2DM that can reach into all geographical and socioeconomic areas of Australia – this is especially important for non-metropolitan communities where area-level SES circumstance tends to be lower [35]. The PDST addressed Goal 2 of the Australian National Diabetes Strategy [14] (i.e., to promote awareness and early detection), in that it enabled opportunistic screening of individuals during routine encounters with the community pharmacy in a previously undiagnosed population. The process evaluation provided information on some factors that influence the implementation and sustainability of the service as well as perspectives and experiences of the small proportion of consumers and GPs who responded to our surveys. Our study supports the view that to be successful a screening program must foster positive engagement among consumers, pharmacists, and GPs to reach its true potential in reducing the burden of diabetes. Better integration of pharmacists and GPs in delivery of preventive services in primary care is also needed to ensure that services are not duplicated, and that appropriate continuity of care is achieved.