Objectives: To systematically review the evidence base for a systems approach to healthcare design, delivery or improvement. Design: Systematic review with meta-analyses. Methods: Included were studies in any patients, in any healthcare setting where a systems approach was compared with usual care which reported quantitative results for any outcomes for both groups. We searched Medline, Embase, HMIC, Health Business Elite, Web of Science, Scopus, PsycINFO and CINAHL from inception to 28 May 2019 for relevant studies. These were screened, and data extracted independently and in duplicate. Study outcomes were stratified by study design and whether they reported patient and/or service outcomes. Meta-analysis was conducted with Revman software V.5.3 using ORs—heterogeneity was assessed using I ² statistics. Results: Of 11 405 records 35 studies were included, of which 28 (80%) were before-and-after design only, five were both before-and-after and concurrent design, and two were randomised controlled trials (RCTs). There was heterogeneity of interventions and wide variation in reported outcome types. Almost all results showed health improvement where systems approaches were used. Study quality varied widely. Exploratory meta-analysis of these suggested favourable effects on both patient outcomes (n=14, OR=0.52 (95% CI 0.38 to 0.71) I ² =91%), and service outcomes (n=18, OR=0.40 (95% CI 0.31 to 0.52) I ² =97%). Conclusions: This study suggests that a systems approaches to healthcare design and delivery results in a statistically significant improvement to both patient and service outcomes. However, better quality studies, particularly RCTs are needed. PROSPERO registration number CRD42017065920.

ML algorithms are used to develop prognostic and diagnostic models and so to support clinical decision-making. This study uses eight supervised ML algorithms to predict the need for intensive care, intubation, and mortality risk for COVID-19 patients. The study uses two datasets: (1) patient demographics and clinical data (n = 11,712), and (2) patient demographics, clinical data, and blood test results (n = 602) for developing the prediction models, understanding the most significant features, and comparing the performances of eight different ML algorithms. Experimental findings showed that all prognostic prediction models reported an AUROC value of over 0.92, in which extra tree and CatBoost classifiers were often outperformed (AUROC over 0.94). The findings revealed that the features of C-reactive protein, the ratio of lymphocytes, lactic acid, and serum calcium have a substantial impact on COVID-19 prognostic predictions. This study provides evidence of the value of tree-based supervised ML algorithms for predicting prognosis in health care.

Quality problem or issue: A number of challenges have been identified with current risk assessment practice in hospitals, including: a lack of consultation with a sufficiently wide group of stakeholders; a lack of consistency and transparency; and insufficient risk assessment guidance. Consequently, risk assessment may not be fully effective as a means to ensure safety.Initial assessment: We used a V system developmental model, in conjunction with mixed methods, including interviews and document analysis to identify user needs and requirements.Choice of solution: One way to address current challenges is through providing good guidance on the fundamental aspects of risk assessment. We designed a risk assessment framework, comprising: a risk assessment model that depicts the main risk assessment steps; risk assessment explanation cards that provide prompts to help apply each step; and a risk assessment form that helps to systematize the risk assessment and document the findings.Implementation: We conducted multiple group discussions to pilot the framework through the use of a representative scenario and used our findings for the user evaluation.Evaluation: User evaluation was conducted with 10 participants through interviews and showed promising results.Lessons learned: While the framework was recommended for use in practice, it was also proposed that it be adopted as a training tool. With its use in risk assessment, we anticipate that risk assessments would lead to more effective decisions being made and more appropriate actions being taken to minimize risks. Consequently, the quality and safety of care delivered could be improved.