Introduction

Improving health care performance is a critical goal for healthcare systems worldwide [1]. Healthcare organizations may improve their performance through clinicians engaging in research activities with research institutions. This assumption has led to policies that promote clinicians’ research engagement to improve health care performance [2–5].

There are several mechanisms of performance improvement at individual and institutional levels. Conducting research activities can develop research networks [2] and enhance clinicians’ absorptive capacity—healthcare providers’ willingness to implement the evidence-based practices generated from their research [5]. Research activities can also include participating in trials; doctors in trial-participating hospitals change their prescription practices quicker, and in-hospital mortality in these hospitals is lower in some departments due to the “trial effect” [4]. Incidentally, the beneficial effects of clinicians’ research activities are reported in cases of myocardial infarction [6–8], an area in which many randomized clinical trials are conducted [9].

Various research activities have the potential to improve health care performance, but the effectiveness of other research activities remains unexplored. Scholarly pursuits, a type of research activity, are mandatory for residents in many countries [10, 11], and presenting at conferences is the first stage of scholarly pursuits in Japan. Presenting research at conferences may promote the characteristics in clinicians that lead to improved health care performance at individual and institutional levels. However, there are few studies specifically investigating the relationship between conference presentations given by hospital physicians and the outcomes of patient care at an institutional level.

Our hypothesis is that clinicians’ conference presentations might be associated with improved patient outcomes through increased evidence-based practice. Therefore, we examined the association between clinicians’ conference presentations and mortality in acute myocardial infarction (AMI) cases after adjusting for potentially confounding patient characteristics. We further confirmed that the proportions of prescribed medical therapies were optimal according to treatment guidelines for evidence-based practice.

Materials and methods

Results

Characteristics of study population

We identified a total of 60,567 patients in 384 acute care hospitals admitted due to acute myocardial infarction. The flow chart for patient selection is illustrated in Fig 1. We excluded patients with missing Killip classification (n = 3,474) and ambulance use (n = 70) data. Thus, 56,923 patients were ultimately included in the complete case analysis (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Patient flow chart.

https://doi.org/10.1371/journal.pone.0315217.g001

Patient characteristics and prescriptions of optimal medical therapies

Patient characteristics are shown in Table 1. As indicated, in-hospital mortality was 7.7% in the Presentation group and 10.9% in the Control group. The number of cases in which drugs were prescribed after admission was significantly greater in the Presentation group than in the Control group for all drugs considered in the study (p-value: < .001).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Demographic and clinical characteristics of the populations.

https://doi.org/10.1371/journal.pone.0315217.t001

Multivariable logistic regression analysis for in-hospital mortality

Table 2 shows the results of the multivariable analysis of in-hospital mortality after adjusting for the covariates in the three models. As indicated, clinicians’ conference presentations were significantly associated with lower in-hospital mortality in all models, except Model 4 (OR = 0.68, 95% CIs: 0.65 to 0.72 in the Crude model; OR = 0.73, 95% CIs: 0.68 to 0.79 in Model 1; OR = 0.76, 95% CIs: 0.70 to 0.82 in Model 2; OR = 0.84, 95% CIs: 0.76 to 0.92 in Model 3; OR = 1.00, 95% CIs: 0.92 to 1.09 in Model 4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Odds ratios for in-hospital mortality associated with conference presentation.

https://doi.org/10.1371/journal.pone.0315217.t002

Discussion

In our analysis using an administrative database that included data from 384 acute care hospitals, a larger number of conference presentations was significantly associated with lower in-hospital mortality in all multivariate logistic regression models used, and this association was also confirmed in sensitivity analyses. In addition, more patients admitted in the Presentation group tended to receive optimal medical therapies, according to treatment guidelines and/or evidence-based practice, compared with patients in the Control group.

Among the strengths of our study is its inclusion of the body of evidence in the cardiovascular field and its treatment of Killip classification as a confounding factor. The cardiovascular department has accumulated a body of evidence based on many randomized clinical trials [9], and greater adherence to treatment guidelines may bring about the lower mortality shown in our results. In our analysis, the c-statistic for Model 2 was 0.91 (95% CI: 0.91–0.92), meaning good discrimination, mainly because the important prognostic factors for AMI, including Killip classification, were included. Actually, several researchers have shown Killip classification to be a good prognostic factor [17, 18]. Similarly, conference presentations may be associated with lower in-hospital mortality in other departments, if those departments acted according to evidence-based practices and established clinical guidelines.

The study described here revealed an association between the number of conference presentations and patient mortality in cases of AMI. Past reports have shown that institutional or operator factors may influence the mortality of AMI [19–21], partially because conference presentations might be related to these associations. In multiple acute care hospitals, the presence of more board-certified cardiologists and lower institutional volume was associated with a lower risk of in-hospital mortality among patients with cardiovascular disease [20, 21]. On the other hand, the annual operator volume was not related to in-hospital mortality [20]. Once the JCS has certified a doctor as a cardiologist in Japan, these doctors must make presentations at JCS conferences to become cardiologists. Similarly, The Accreditation Council for Graduate Medical Education (ACGME) Common Program mandates faculty scholarly activities with an active research component and supports residents in their scholarly activities [11], while the Japanese Society of Internal Medicine imposes on residents the requirement of performing at least two scholarly activities in order to become internal medicine specialists. Plausibly, conference presentations may be an important mechanism to ensure the quality of medical care and an associated lower risk of in-hospital mortality.

Several studies have shown that scholarly activities by hospital physicians may lead to lower mortality, greater adherence to treatment guidelines and/or an increase in evidence-based practice [4, 7, 8]. Jha et al. studied the survival of AMI patients in hospitals conducting clinical trials and found that the survival rate for participants was higher than the survival rate for non-participants [8]. Similarly, Majumdar et al. reported that patients treated in hospitals whose physicians participated in trials showed significantly lower mortality than those treated in non-participating hospitals [7]. In this previous literature, a composite guideline adherence score was used and the uptake of nine recommendations from the guidelines was assessed. It was found that the adherence proportion tended to increase significantly with the participation level. In our analyses, the patients in the Presentation group were consistently prescribed the five evidence-based drugs in more cases than those admitted to hospitals with no annual conference presentations. The difference in mortality in our results may be explained by these prescriptions.

There are several limitations to the study. First, “conference presentation” was defined as a presentation only at the annual scientific meetings of the JCS, which is recognized as one of the major medical societies in Japan in view of the importance of cardiovascular diseases. (As of June 2021, the JCS has 26,970 regular members, 4,570 associate members, and 15,098 cardiovascular specialists.) We have chosen the JCS conference based on the current healthcare performance, but it is fully recognized that similar effects could be achieved with another conference presentation, such as those held by the Japanese Society of Internal Medicine or the Japanese Association of Cardiovascular Intervention and Therapeutics. Second, unmeasured confounding hospital and personal factors exist, such as organizational culture, the cardiologist’s motivation, and NYHA classification. Future observational studies with comprehensive measurements of these confounding factors would enhance our understanding of their effects on patient mortality. Thus, our cohort study might feasibly provide evidence that can be considered somewhat credible.

Conclusion

We concluded that conference presentations were associated with lower in-hospital mortality, and patients tend to benefit from more evidence-based practice in hospitals whose physicians make conference presentations.

Acknowledgments

The authors wish to thank all the staff members and all the participants in the hospitals participating in the QIP project.