Article Text
PDF
Statistics from Altmetric.com
Introduction
Health systems invest in diabetes quality improvement (QI) programmes to reduce the gap between research evidence of optimal care and current care.1 Examples of commonly used QI strategies in diabetes include programmes to measure and report quality of care (ie, audit and feedback initiatives), implementation of clinician and patient education, and reminder systems. A recent systematic review of randomised trials of QI programmes indicates that they can successfully improve quality of diabetes care and patient outcomes.2 Changes in surrogate markers such as blood glucose control, blood pressure or cholesterol levels are used to measure QI intervention effectiveness.2
However, investments in QI strategies are only worthwhile if the programmes that effectively improve care are sustained after trial completion.3 Failure to maintain QI programmes contributes to substantial research waste, resulting in suboptimal patient care since the effective interventions are not available.4 5 Furthermore, failure to redirect resources from ineffective programmes creates opportunity cost. To date, no studies have examined the sustainability of rigorously evaluated diabetes QI programmes. The objective of this study is to explore factors associated with sustained implementation of diabetes QI programmes after cessation of their research funding.
Methods
In 2018, we emailed the authors of 226 trials on diabetes QI programmes and requested them to complete an online survey about their perceived sustainability of their intervention. These trials were published between 2004 and 2014, and were identified in a systematic review.1 A follow-up email and phone calls were used to prompt survey completion. This study received ethics exemption.
Our main outcome of interest was whether the QI intervention was sustained in the study settings after the trial was completed. Study characteristics collected from trial authors included funding source (local funding vs others), whether the funding covered both intervention implementation and evaluation or evaluation alone, and whether the intervention improved patient care. Sample size, year of publication, country, duration of intervention and follow-up, and whether the study was conducted at single or multiple sites were extracted from the study publications. Chi-square tests were used to assess the association between study characteristics and the sustained local implementation of the QI interventions.
Results
Ninety-four of the 226 authors (42%) responded to the survey. As shown in table 1, authors in the United States had a higher response rate compared with those from other countries. Studies conducted by survey responders and non-responders were similar with respect to year of publication, sample size, study duration, number of QI strategies included and reported effect size of the QI interventions.
Study characteristics and effect sizes of quality improvement interventions among survey responders and non-responders
Study characteristics, stratified by whether the intervention was sustained after the trial, are shown in table 2. 78% (73/94) of trials reported improved quality of care, but 40% (29/73) of these trials were not sustained following study completion. Surprisingly, QI programmes were sustained in 19% (4/21) of studies, where no improvement in quality of care was achieved. Of the variables examined, only improvement of care was significantly associated with sustained local implementation of interventions (p=0.002). Funding source, funding coverage, single versus multiple study sites and study duration were not significantly associated with sustained implementation.
Factors associated with sustained implementation of diabetes quality improvement interventions at trial clinical settings
We classified the QI strategies into 12 categories using a previously described framework.2 Online supplementary table 1 presents the strategies used, which are categorised based on whether the interventions were sustained. Most trials included more than one QI strategy for a median of 3.5 strategies for sustained interventions and three for those that were not sustained.
Supplemental material
Discussion
This is the first study to explore whether diabetes QI interventions evaluated in trials were sustained after completion of the trial. Although half of the responding authors reported that the interventions assessed in their trials were sustained in local settings, 40% of interventions deemed effective were not sustained. This finding suggests that patients may not be benefitting from the research. 19% of interventions deemed ineffective were sustained, thereby wasting healthcare resources. Further work is required to explore why these ineffective interventions were sustained and if unreported changes to the interventions after trial completion may play a role.
This study also represents the first empirical analysis of variables that may affect the sustained implementation of diabetes QI interventions across a wide range of settings. Although we assessed several characteristics to determine their impact on the sustained implementation of interventions locally, only improvement of care was significantly associated with this outcome. Contrary to our expectations, nature of the funding was not associated with sustainability. Limitations of these findings include the sample size and a response rate of 42%. Non-responders may have had less sustained implementation of their interventions, so the current results may overestimate the proportion of diabetes QI interventions that are sustained. As this study explores perceived sustainability based on survey responses, qualitative work is underway to further investigate how the authors concluded whether their intervention was sustained. Online supplementary table 1 indicates some QI strategies, such as audit and feedback, that may be more likely to be sustained. Further work is needed to explore the sustainability potential of various QI strategies, including how to support the integration of these strategies into routine practice, how to consider sustainability from the beginning of a QI intervention and the use of existing theories and frameworks to support sustainable implementation.
Interventions shown to improve quality of diabetes care are not reliably sustained after the trial is completed. Since optimal diabetes care generally requires intensive management involving multidisciplinary teams,2 effective interventions that support high-quality care cannot be wasted. Sustainability of effective interventions requires local healthcare systems to take responsibility for the resources used by study authors and consider the long-term needs of diabetes management. Those conducting diabetes QI initiatives should engage with the healthcare system and patients at the beginning of any QI study to ensure that the project is aligned with its needs, to confirm capacity for sustained implementation, and to agree on the outcomes to inform sustainability.
Footnotes
ELK-H and HYY are joint first authors.
ELK-H and HYY contributed equally.
Contributors NMI has full access to all the data in the study and takes responsibility for the integrity of the data and accuracy of the data analysis. Study concept and design: JMG and NMI. Acquisition of data: ELK-H and HYY. Analysis and interpretation of data: ELK-H, HYY, SK, JG, and NI. Drafting of the manuscript: ELK-H and HYY. Critical revision of the article for important intellectual content: ELK-H, HYY, SK, CL, JG, and NMI. Statistical analysis: SK. Study supervision: JMG and NMI.
Funding JMG holds a Canada Research Chair in Health Knowledge Transfer and Uptake. NMI holds a Canada Research Chair in Implementation of Evidence Based Practice.
Competing interests None declared.
Patient consent for publication Not required.
Provenance and peer review Not commissioned; externally peer reviewed.