Recruiting family physicians or primary care practices into research studies is daunting. The evidence base supporting primary care recruitment is slowly increasing and highlighting the many steps, significant resources, and considerable time required to involve community-based health care providers in research (Asch et al., Reference Asch, Connor, Hamilton and Fox2000; Goodyear-Smith et al., Reference Goodyear-Smith, York, Petousis-Harris, Turner, Copp, Kerse and Grant2009; Jones et al., Reference Jones, Dixon, Falkingham, Piteman and Dixon2011; McKinn et al., Reference McKinn, Bonner, Jansen and McCaffery2015). Recruiting primary care study participants often requires ongoing communication to identify correct contact information, confirm eligibility, share study information and elicit an expression of interest, share and ensure return of study documents, and carry out engagement and retention initiatives over the duration of the study (Johnston et al., Reference Johnston, Liddy, Hogg, Donskov, Russell and Gyorfi-Dyke2010; Horspool et al., Reference Horspool, Julious, Mooney, May, Sully and Smithson2015). Many of these steps are multiplied by more doctors or practices than are needed for a study sample. Despite the growing body of evidence on recruitment approaches for primary care, strategies to facilitate recruitment and manage the ongoing information flow necessary for large community-based samples of independent practices remains a persistent gap in the collective knowledge base.
Our team sought to improve the management of the information needed for participant recruitment over a three- to six-month period for a geographically dispersed study of primary care performance in three regions across Canada in Ontario, Nova Scotia, and British Columbia. We had previous experience using Microsoft Excel and Access (Microsoft Office Professional Plus 2013) for recruitment tracking. These programs are inexpensive and easy to learn but remain basic data storage tools, not information management tools, and they do not support any automation, linking, or reminder systems to manage and integrate recruitment information and processes.
This report shares some of our lessons learned in experimenting with a customer relationship management (CRM) program – designed to improve customer relations and communication in business and organizations – to make the recruitment, management, and retention of physician and practice participants more effective and efficient.
CRM program process
Our study had project coordinators in three regions actively recruiting and managing participants over a two-year period with a goal of 40 physicians located in 40 different practices in each region. We needed a system to:
∙ Catalogue all potentially eligible practices for recruitment in the regions;
∙ Catalogue practice information including criteria for study eligibility, contacts (eg, receptionist, practice manager, etc.), and providers;
∙ Catalogue the recruitment efforts at each step – from initial contact to follow-up attempts, to sending and following up on study documents;
∙ Record practice preferences for when to follow-up, who to contact and personal relationship with investigator;
∙ Remind when follow-ups for specific practices were due (eg, Practice X asked for a call the first week of September, Practice Y needs a follow-up call in two weeks);
∙ Track when study documents (eg, consent forms, completed surveys, etc.) were received or if still outstanding;
∙ Schedule site visits for data collection.
We considered 12 CRM programs listed in at least two separate rankings in top 10 independent online reviews of such programs. The software reviews compared user-friendliness, customer support, feature customization, integration with Microsoft Office and Outlook (Microsoft Office Professional Plus 2013) products, regulatory compliance, cost, analysis tools, and minimum contract duration.
The web-based Workbooks CRM (Workbooks Online Ltd., 2016) software system was chosen for high ratings of user-friendliness, customer support, customizability, and cost. Attractive features included real-time updates, comprehensive communication tracking (with automatic reminders sent to email or calendar), linkage with email accounts to contact participants directly, and efficient support. However, customized fields had to be manually created, there was no view of all fields at the same time and it was a two-step process to initially enter practice physicians into the system.
We bought the program ($1181.11 Canadian for one year). Site coordinators underwent an initial general training of several hours through the online training support provided with the contract and had a teleconference with the vendor’s support team. The system was then deployed in the three provinces. Coordinators were learning and deploying several new systems at the same time, including an internet-based communication platform, cloud-based reference management software and secure file sharing program, and the recruitment CRM program. Coordinators successfully adopted each of the other new systems. However, after three- to six-months of use, each of the three sites abandoned the CRM program as it did not make the recruitment process easier. In each province, the staff fell back to tried-and-true Microsoft Excel (Microsoft Office Professional Plus 2013) for recruitment data management. We did not use the CRM long enough to assess the impact on retention.
Overall lessons learned
We explored, as a team, why the software had not been successful in any of the three regions. Each of the three site coordinators was interviewed separately, by someone who was not part of the research team, to ensure comprehensive collection of feedback on their experiences using the new software for recruitment. This feedback was recorded, transcribed, and shared with the research team. Site coordinators and principal investigators discussed the feedback to identify, through consensus, the key lessons from our collective experience. All three coordinators consented to their feedback being published.
Our team acknowledged that there is always a certain amount of resistance to new technology and a desire to cling to familiar processes. We expected this and tried to compensate for it by selecting a new product that might better align with our needs. From initiation to training, site coordinators were enthusiastic about learning a potentially more efficient program. The other new systems deployed at the same time were all successfully integrated into ongoing operations signalling the team had capacity for learning new approaches.
Though coordinators who used the CRM software saw it had potential, they expressed that in order for the product to be a successful management tool, additional time for initial training and more accessible ongoing support would have been necessary. Said the Ontario coordinator: ‘It is my belief that this program may be good for a longer study but the ramp up time it would take to learn the software did not equate to the amount of time we would spend recruiting and the number of touches we would have’.
The coordinators unanimously called for more training and better access to support when needed as issues arose with the CRM program. The Nova Scotia coordinator noted: ‘The system has potential but we were unable to actualize the potential given our limited understanding of the system. If someone were to use this system for research recruitment again, it would be really helpful to be thoroughly trained. When we sought help, we were told we would need to pay a consultation fee to get specific feedback about our system, in addition to the existing cost for the subscription. However, it is important to fully understand how to use the system’.
The strengths and challenges encountered with the software for each step of the recruitment process in each region are described in Table 1.
Project coordinators’ views on using the software
Conclusions
We believe new software can – and should – be used to make the collection, collation, and distribution of information during a research study more efficient and effective. However, use of the CRM program was too much of a leap from business to research for the timeframe and requirements of our study. Limited time to learn new technology and lack of comprehensive support forced staff to struggle with balancing, mastering, and using the technology while meeting the recruitment needs of the study. In addition to the cost of purchasing such a software product, a research team would need to budget for a significant amount of training time and ongoing technical support from the product developers to ensure the team members using the product were able to do so efficiently and effectively. If multiple new systems need to be learned at the same time, a coordinated strategy should recognize the significant time investment required for mastering each of these.
We found improving recruitment is not simply squeezing our needs into a CRM program meant for other things. Software and support packages better adapted to meet the needs of primary care health care research should be sought to minimize burdening the project staff with customization. Most projects cannot afford long learn-times and tighter recruitment timelines need rapid support on demand. We never reached the stage of time saving where automatic mail outs or reminders to follow-up might have reduced the burden on recruitment coordinators. Regardless, staff could see the potential benefit for longer studies capable of investing more time upfront in proper training. This might be especially attractive to long-term research programs or projects involving the same sample of potential participants over multiple studies. In the meantime, there is still a need for more accessible recruitment management programs for the myriad studies in primary care research.
Acknowledgements
The authors wish to acknowledge the valuable contributions of Jackie Schultz throughout the project as well as the work of Ottawa writer/editor Joan Ramsay for her help in developing this article.
Financial Support
This work was supported by the Canadian Institutes of Health Research (grant number TTF-128265); and the Michael Smith Foundation for Health Research (grant number PT-CPH-00001-134).
Conflicts of Interest
None.
