Domhnall MacAuley is a CMAJ Associate Editor and a professor of primary care in Northern Ireland, UK.
Research conferences should be an opportunity to gain insights from discussion and collegial debate about new research. At times, though, I have seen debate become adversarial and counterproductive; questions can be aggressive and speakers defensive. But one of the great attributes of the North American Primary Care Research Group annual meeting (NAPCRG) is the culture intellectual rigor yet respectful and collegial discussion, and the support for early career researchers. Researchers with impressive track records in publication in international journals are always keen to share their knowledge and help their colleagues. David Meyers, a long time NAPCRG supporter unable to attend the conference this year due to illness sent a video message in which he said, “May you find meaning in your work and friendship in your colleagues.”
Michael Moore, presented a novel study with great potential to radically change practice. He and colleagues wanted to compare the effect of using analgesic drops alone with usual care and placebo drops for children with ear ache. Their randomized controlled trial, although elegantly designed, did not reach definitive conclusions. Delay in delivery of the placebo drops and deficits in recruitment hampered the study’s progress. The trial was terminated early by the funders and the results were insufficient to meet statistical thresholds; however, they showed a strong signal in favour of using analgesic drops alone, with a reduction in antibiotic prescriptions. Clinical research can be imperfect, even when impeccably designed. This great study deserves to be redone.
Moore presented more impressive research in a later distinguished paper keynote address. Is there a difference between high and low prescribing practices in terms of types of patients seen. Moore and colleagues reanalyzed data from three enormous prospective studies with prescribing data from almost 1000 practices. There was considerable variability in the prescribing rate in each of the data sets. Higher prescribing practices do see more ill patients, apparently. Although this does not explain the entire effect, it suggests that it may be inappropriate to apply simplistic prescribing guidelines across all practices; targets should be in the context of illness.
Every clinical contact can generate research questions and Kome Gbinigle’s work was clearly grounded in clinical practice. We are taught in medical school about the symptoms and signs of pneumonia: cough, sputum, chest crackles etc. But, in her experience, older patients did not always fit this clinical picture so she decided to review the literature to find out if that was true. To tailor her research question to her experience in daily practice she decided to exclude studies of patients who were immunosuppressed and those studies that recruited patients from emergency departments as such patients had a high likelihood of serious illness. So she focused on seven highly relevant studies that allowed her to assess the pre- and post-test probability of each clinical sign. She found the classic characteristic of cough, chest pain, sputum and chest crackles were not particularly helpful in diagnosis, but vital signs were. Tachycardia, tachypnea, low BP, and impaired cognitive function were much more helpful.
Mark Ebell, who usually publishes high quality epidemiological work, presented an interesting collaborative scenario study with his Swiss colleagues from Lausanne looking at decision thresholds in diagnosis and treatment of community acquired pneumonia. These scenarios were generated using probability based on GRACE – and they provided doctors with before and after information. It was interesting to see how clinicians significantly overestimated the likelihood of pneumonia. If there was a chest x-ray available on site they were less likely to treat, ED doctors were less likely to treat, and Swiss doctors in the study were less likely to treat empirically. But, where Ebell’s expertise really came into play was in interpreting the threshold diagrams and in determining the threshold levels for each component of the clinical vignette. The key message was that if you develop a clinical support tool you can marry it to a test threshold model, in which a low risk threshold should be low enough and a high risk threshold high enough.
Kurt Stange, who was recognized for his lifetime contribution to primary care research, highlighted some key aspects of research and practice in his award acceptance address. For example, he described how family medicine is about balancing the biological and the biographical. Illnesses do not occur in isolation – within a family medicine diabetic follow up consultation he could identify 25 separate issues ranging from the psychosocial to the physiological. There are challenges for the family doctor in prioritizing these problems. As a fellow medical editor I also recognize Kurt’s immense contribution in his 11 years as EIC of Annals of Family Medicine.
In the academic equivalent of celebrity spotting, I found myself sitting beside the author of the definitive paper on multimorbidity, Susan Smith, as speaker after speaker, in a session devoted to multimorbidity, cited her work. I wondered if they knew she was in the audience. Maxime Sasseville presented a qualitative study looking at seven of the nine key domains in multimorbidity, a thematic analysis that identified the movement of patient power in consultation, self-efficacy, more self confidence in care, and knowledge acquisition. Katie Gallacher looked at treatment burden and patient capacity – the ability of patients to manage their health. This study included both hospital and community health professionals and asked about facilitators and barriers to health care. The barriers were particularly difficult discharge transition, especially highlighting how medical and social care systems struggled to communicate with each other and noting lack of transport services.
Some research presentations at conferences just stand out and one such on multimorbidity intervention by Chris Salisbury was exceptional. Researchers have defined the patterns, prevalence, and various components of multimorbidity and their implications in many different populations, but research appeared to have stalled on interventions. Few researchers have successfully trialled multimorbidity interventions. However, in their patient-centred and elegantly designed randomized controlled trial of a patient focused approach aimed at improving outcomes in patients with multimorbidity, Salisbury and colleagues found that the intervention made little difference compared with usual care – an important, albeit negative finding.
Sander van Doorn looked at the potential biomarker NTproBNP in screening for atrial fibrillation in primary care. He gave us his conclusions at the outset, stating simply: Don’t use this biomarker to identify atrial fibrillation. But, don’t let that straightforward conclusion mislead you as to the complexity of his work of analyzing data from four large practice-based screening studies. Using cut off points of 125pg/ml and 400pg/ml, he determined the sensitivity and specificity, predictive values, proportion below cut off and prevalence of heart failure below the cut point. Nearly half of their patients with atrial fibrillation also had heart failure. The test would be positive too often and negative too often and an echo would be needed in every patient identified with atrial fibrillation.
Felicity Goodyear Smith presented a study in which global stakeholders were asked to identify the research gaps in low- and middle-income countries, a key interest for the world organization of family doctors (WONCA). As might have been expected, this generated many questions and particularly those related to health system reform.
Beyond the presentation of research findings, NAPCRG is an opportunity to develop national and international collaborations. There was a large international presence and, while English may be the first language of scientific communication, there was a large Francophone group at the conference from Quebec, Canada, and from across Europe. We invited some of these colleagues to give their reflections on camera. This also includes an explanation of one of the most talked about posters at the conference- the Dragon’s Den innovation which caught everyone’s imagination.
Encouraging medical students to follow a career in primary care, let alone a research track, is a major challenge. This appears especially difficult In the US where students have accrued so much debt that, although they have been inspired to follow a career in family medicine, they may take a pragmatic approach to their careers that is determined by the financial imperative of managing their debt. With these pressures, they may have little option but to opt for high remuneration specialties. It is also difficult to attract generalists – either in general internal medicine or in family medicine – as medicine has become so focused that neither patients nor doctors are happy to tolerate uncertainty. Academic primary care is a relatively small interest group. While NAPCRG attracted several hundred delegates, the American Heart Association meeting had upwards of 15,000 delegates taking place at the same time. The major influential factors in recruiting and retaining doctors may vary but include a matrix of income, workload, professional esteem, the need for peer support, gender changes in the workforce, and positive exposure in undergraduate study. In this context it was encouraging to me to meet several medical students attending the NAPCRG meeting, who were enthusiastic about research and had already undertaken some studies. Let them speak from themselves:
NAPCRG 2018’s final keynote address was delivered Bernard Ewigman who took us through his personal journey in his research career. I agreed with him about the importance of capacity building and was surprised at his assessment of the dearth of established primary care research centres in the United States. At a later session, more comprehensive data were presented from the CERA-FMAH-ADFM study that confirmed this. In describing the pivotal moments in his research career, Ewigman made a comment that caught my imagination – that the most important question that any mentor should ask a nascent researcher, faced with a seemingly impossible research study, is, “Why not?”