A hybrid Shewhart chart for visualizing and learning from epidemic data.

OBJECTIVE: As the globe endures the coronavirus disease 2019 (COVID-19) pandemic, we developed a hybrid Shewhart chart to visualize and learn from day-to-day variation in a variety of epidemic measures over time. CONTEXT: Countries and localities have reported daily data representing the progression of COVID-19 conditions and measures, with trajectories mapping along the classic epidemiological curve. Settings have experienced different patterns over time within the epidemic: pre-exponential growth, exponential growth, plateau or descent and/ or low counts after descent. Decision-makers need a reliable method for rapidly detecting transitions in epidemic measures, informing curtailment strategies and learning from actions taken. METHODS: We designed a hybrid Shewhart chart describing four 'epochs' ((i) pre-exponential growth, (ii) exponential growth, (iii) plateau or descent and (iv) stability after descent) of the COVID-19 epidemic that emerged by incorporating a C-chart and I-chart with a log-regression slope. We developed and tested the hybrid chart using international data at the country, regional and local levels with measures including cases, hospitalizations and deaths with guidance from local subject-matter experts. RESULTS: The hybrid chart effectively and rapidly signaled the occurrence of each of the four epochs. In the UK, a signal that COVID-19 deaths moved into exponential growth occurred on 17 September, 44 days prior to the announcement of a large-scale lockdown. In California, USA, signals detecting increases in COVID-19 cases at the county level were detected in December 2020 prior to statewide stay-at-home orders, with declines detected in the weeks following. In Ireland, in December 2020, the hybrid chart detected increases in COVID-19 cases, followed by hospitalizations, intensive care unit admissions and deaths. Following national restrictions in late December, a similar sequence of reductions in the measures was detected in January and February 2021. CONCLUSIONS: The Shewhart hybrid chart is a valuable tool for rapidly generating learning from data in close to real time. When used by subject-matter experts, the chart can guide actionable policy and local decision-making earlier than when action is likely to be taken without it.

Understanding variation in reported covid-19 deaths with a novel Shewhart chart application.

OBJECTIVE: Motivated by the coronavirus disease 2019 (covid-19) pandemic, we developed a novel Shewhart chart to visualize and learn from variation in reported deaths in an epidemic. CONTEXT: Without a method to understand if a day-to-day variation in outcomes may be attributed to meaningful signals of change-rather than variability we would expect-care providers, improvement leaders, policy-makers, and the public will struggle to recognize if epidemic conditions are improving. METHODS: We developed a novel hybrid C-chart and I-chart to detect within a geographic area the start and end of exponential growth in reported deaths. Reported deaths were the unit of analysis owing to erratic reporting of cases from variability in local testing strategies. We used simulation and case studies to assess chart performance and define technical parameters. This approach also applies to other critical measures related to a pandemic when high-quality data are available. CONCLUSIONS: The hybrid chart detected the start of exponential growth and identified early signals that the growth phase was ending. During a pandemic, timely reliable signals that an epidemic is waxing or waning may have mortal implications. This novel chart offers a practical tool, accessible to system leaders and frontline teams, to visualize and learn from daily reported deaths during an epidemic. Without Shewhart charts and, more broadly, a theory of variation in our epidemiological arsenal, we lack a scientific method for a real-time assessment of local conditions. Shewhart charts should become a standard method for learning from data in the context of a pandemic or epidemic.

A healthy bottom line: healthy life expectancy as an outcome measure for health improvement efforts.

CONTEXT: Good health is the most important outcome of health care, and healthy life expectancy (HLE), an intuitive and meaningful summary measure combining the length and quality of life, has become a standard in the world for measuring population health. METHODS: This article critically reviews the literature and practices around the world for measuring and improving HLE and synthesizes that information as a basis for recommendations for the adoption and adaptation of HLE as an outcome measure in the United States. FINDINGS: This article makes the case for adoption of HLE as an outcome measure at the national, state, community, and health care system levels in the United States to compare the effectiveness of alternative practices, evaluate disparities, and guide resource allocation. CONCLUSIONS: HLE is a clear, consistent, and important population health outcome measure that can enable informed judgments about value for investments in health care.

Learning networks for sustainable, large-scale improvement.

Large-scale improvement efforts known as improvement networks offer structured opportunities for exchange of information and insights into the adaptation of clinical protocols to a variety of settings.

A poisson-based prediction model and warning system for MRSA daily burden.

OBJECTIVE: This study was designed to demonstrate that the number of methicillin-resistant Staphylococcus aureus (MRSA) isolates collected daily in a community hospital is Poisson distributed and that using a one-sided Poisson control table is a fast and easy way to recognize unusually high numbers of MRSA isolates collected daily that may signal possible outbreaks. METHODS: A retrospective analysis of MRSA isolates collected daily over a three year period (2005-2007, N = 934) was performed. Observed MRSA isolate frequencies are compared to Poisson frequencies using chi-square goodness-of-fit tests. A regression equation on the mean number of MRSA isolates collected daily for the years 2005, 2006, and 2007 is used to predict the mean number of MRSA isolates for 2008. A warning system for MRSA isolates collected daily is presented and a one-tailed, mean + 2 sigma control table is provided. SETTING: One-hundred-fifty bed community hospital in central Massachusetts. RESULTS: Goodness-of-fit tests showed close agreement between actual MRSA isolates collected daily and Poisson frequencies for 2005 (chi4(2) = 4.045, p = 0.39), 2006 (chi4(2) = 2.807, p = 0.59), and 2007 (chi4(2) = 1.494, p = 0.83). CONCLUSION: Theoretical and empirical support is provided for the Poisson probability model. The model can be used to identify unusually high occurrences ofMRSA isolates collected daily. This study was limited to a single community healthcare system but the results may be generalized to other types of healthcare settings.

A long-term forecast of MRSA daily burden using logistic modeling.

OBJECTIVE: This article presents a logistic model that describes the mean number of unique methicillin-resistant Staphylococcus aureus (MRSA) isolates collected daily at a 150-bed community hospital in central Massachusetts. The model is used to derive a long-term forecast of the mean MRSA isolate frequency. METHODS: The mean number of MRSA isolates collected daily was found for each quarter from the first quarter of 1996 to the first quarter of 2008. A logistic model was fit to the data and then extrapolated to obtain a long-term forecast. SETTING: Data was collected at a one-hundred-fifty bed community hospital in central Massachusetts. RESULTS: The coefficient of determination indicates that 87% of the variation in transformed data is explained by the model. The extrapolated logistic model prediction is that the mean number of MRSA isolates collected daily approaches 1.42 MRSA isolates per day. CONCLUSION: Logistic modeling of empirical data using modest mathematical assumptions is an effective way to understand, visualize, and forecast MRSA daily frequencies over time. The advantage for laboratorians and epidemiologists is that logistic models provide reliable trending and long-term prediction ability of multi-drug resistant organism frequencies. Moreover, as additional data is obtained, the logistic model assumptions can be checked, the model updated, and forecasts improved.

Integrating Social Needs Into Health Care: A Twenty-Year Case Study Of Adaptation And Diffusion.

The US health care system has recently begun to account for patients' unmet social needs in care delivery and payment reform. This article presents a twenty-year qualitative case study of five stages of diffusion-testing and learning, standardization, replication, shifting from doing to enabling, and catalyzing broad adoption-of a practical approach for integrating social needs into clinical care. This case study of Health Leads and its funders confirms the importance of focusing on a clear aim, investing in model testing and standardization to enable subsequent responsiveness to the market, and the willingness of innovators and their investors to cede control of a model to allow local adaption and accelerate broad adoption.

Government As Innovation Catalyst: Lessons From The Early Center For Medicare And Medicaid Innovation Models.

Congress established the Center for Medicare and Medicaid Innovation (CMMI) to design, test, and spread innovative payment and service delivery models that either reduce spending without reducing the quality of care or improve the quality of care without increasing spending. CMMI sought to leverage these models to foster market innovation and accelerate the transformation of payment and care delivery to achieve the Triple Aim of better health, better care, and lower cost. This article provides a perspective on the design and execution of CMMI's five initial models, the resulting outcomes and lessons, and how their core concepts evolved within and spread beyond CMMI. This experience yields three key insights that could inform future efforts by CMMI and public and private payers, including model designs and policy decisions. These insights center on the need for iterative testing and learning guided by market feedback, more realistic time frames to demonstrate impact on cost and quality, and greater integration of models.

Measuring Hospital-Wide Mortality-Pitfalls and Potential.

Risk-adjusted hospital-wide mortality has been proposed as a key indicator of system-level quality. Several risk-adjusted measures are available, and one-the hospital standardized mortality ratio (HSMR) - is publicly reported in a number of countries, but not in the United States. This paper reviews potential uses of such measures. We conclude that available methods are not suitable for interhospital comparisons or rankings and should not be used for pay-for-performance or value-based purchasing/payment. Hospital-wide mortality is a relatively imprecise, crude measure of quality, but disaggregation into condition- and service-line-specific mortality can facilitate targeted improvement efforts. If tracked over time, both observed and expected mortality rates should be monitored to ensure that apparent improvement is not due to increasing expected mortality, which could reflect changes in case mix or coding. Risk-adjusted mortality can be used as an initial signal that a hospital's mortality rate is significantly higher than statistically expected, prompting further inquiry.

The "physician-led chart audit: " engaging providers in fortifying a culture of safety.

OBJECTIVES: Engaging physicians in quality and patient safety initiatives is a well-described challenge. Barriers include time constraints, lack of defined common purpose and leadership support, poorly communicated goals, and scarcity of supporting data (references 1, 2, 3).With reference to strengthening a culture of safety while meeting regulatory and performance standards, health-care systems face a difficult twin objective: educate the medical staff and its trainees and maintain high levels of compliance across inpatient, ambulatory clinic, and procedural areas.In 2010, our institution identified opportunities for improvement in physician performance related to several important patient safety standards. These issues had not been previously corrected by didactic sessions, written communication, and the chain of command. METHODS: To help address these general and site-specific problems, we developed and deployed a set of medical record audit tools entitled "physician-led audits" (PLAs). We trained providers in leadership positions to use the tools and to teach their use to others. We designed a system to capture data on frequency of use and results. The PLA process is distinctive in that it holds physicians accountable for the auditing and for follow-up afterward. RESULTS: With support from department chairs, division chiefs, residency program directors, and other leaders, close to 2000 PLAs were performed over a 10-month study period. CONCLUSIONS: The audits engaged physician leaders and the at-large medical staff, making them key participants in a system-wide improvement campaign. The tool was customized for use in a broad range of clinical settings and was widely and rapidly adopted, leading to important dialogue and a substantive contribution to our safety culture.

Sampling considerations for health care improvement.

Sampling in improvement work can pose challenges. How is it different from the sampling strategies many use with research, clinical trials, or regulatory programs? What should improvement teams consider when determining a useful approach to sampling and a useful sample size? The aim of this article is to introduce some of the concepts related to sampling for improvement. We give specific guidance related to determining a useful sample size to a wider health care audience so that it can be applied to improvement projects in hospitals and health systems.

Health care quality improvement publication trends.

To analyze the extent of academic interest in quality improvement (QI) initiatives in medical practice, annual publication trends for the most well-known QI methodologies being used in health care settings were analyzed. A total of 10 key medical- and business-oriented library databases were examined: PubMed, Ovid MEDLINE, EMBASE, CINAHL, PsycINFO, ISI Web of Science, Scopus, the Cochrane Central Register of Controlled Trials, ABI/INFORM, and Business Source Complete. A total of 13 057 articles were identified that discuss at least 1 of 10 well-known QI concepts used in health care contexts, 8645 (66.2%) of which were classified as original research. "Total quality management" was the only methodology to demonstrate a significant decline in publication over time. "Continuous quality improvement" was the most common topic of study across all publication years, whereas articles discussing Lean methodology demonstrated the largest growth in publication volume over the past 2 decades. Health care QI publication volume increased substantially beginning in 1991.

Whole-patient measure of safety: using administrative data to assess the probability of highly undesirable events during hospitalization.

Hospitals often have limited ability to obtain primary clinical data from electronic health records to use in assessing quality and safety. We outline a new model that uses administrative data to gauge the safety of care at the hospital level. The model is based on a set of highly undesirable events (HUEs) defined using administrative data and can be customized to address the priorities and needs of different users. Patients with HUEs were identified using discharge abstracts from July 1, 2008 through June 30, 2010. Diagnoses were classified as HUEs based on the associated present-on-admission status. The 2-year study population comprised more than 6.5 million discharges from 161 hospitals. The proportion of hospitalizations including at least one HUE during the 24-month study period varied greatly among hospitals, with a mean of 7.74% (SD 2.3%) and a range of 13.32% (max, 15.31%; min, 1.99%). The whole-patient measure of safety provides a global measure to use in assessing hospitals with the patient's entire care experience in mind. As administrative and clinical datasets become more consistent, it becomes possible to use administrative data to compare the rates of HUEs across organizations and to identify opportunities for improvement.

Sampling considerations for health care improvement.

Sampling in improvement work can pose challenges. How is it different from the sampling strategies many use with research, clinical trials, or regulatory programs? What should improvement teams consider when determining a useful approach to sampling and a useful sample size? The aim of this article is to introduce some of the concepts related to sampling for improvement. We give specific guidance related to determining a useful sample size to a wider health care audience so that it can be applied to improvement projects in hospitals and health systems.

Seven propositions of the science of improvement: exploring foundations.

CONTEXT: The phrase "Science of Improvement" or "Improvement Science" is commonly used today by a range of people and professions to mean different things, creating confusion to those trying to learn about improvement. In this article, we briefly define the concepts of improvement and science, and review the history of the consideration of "improvement" as a science. METHODS: We trace key concepts and ideas in improvement to their philosophical and theoretical foundation with a focus on Deming's System of Profound Knowledge. We suggest that Deming's system has a firm association with many contemporary and historic philosophic and scientific debates and concepts. With reference to these debates and concepts, we identify 7 propositions that provide the scientific and philosophical foundation for the science of improvement. FINDINGS: A standard view of the science of improvement does not presently exist that is grounded in the philosophical and theoretical basis of the field. The 7 propositions outlined here demonstrate the value of examining the underpinnings of improvement. This is needed to both advance the field and minimize confusion about what the phrase "science of improvement" represents. We argue that advanced scientists of improvement are those who like Deming and Shewhart can integrate ideas, concepts, and models between scientific disciplines for the purpose of developing more robust improvement models, tools, and techniques with a focus on application and problem solving in real world contexts. CONCLUSIONS: The epistemological foundations and theoretical basis of the science of improvement and its reasoning methods need to be critically examined to ensure its continued development and relevance. If improvement efforts and projects in health care are to be characterized under the canon of science, then health care professionals engaged in quality improvement work would benefit from a standard set of core principles, a standard lexicon, and an understanding of the evolution of the science of improvement.

Large-scale improvement initiatives in healthcare: a scan of the literature.

CONTEXT: The goal of this article is to provide a succinct scan of the literature as it relates to the current thinking and practice in large-scale improvement initiatives in healthcare. METHOD: We employed a scan of the literature using a modified Delphi technique. A standard review form was used. The scan was limited to large-scale spread efforts in hospitals and healthcare systems. Each of the main factors that emerged during the scan was linked to secondary factors and organized using a driver diagram. FINDINGS: Four primary drivers (factors) emerged during our scan that inform large-scale change initiatives in healthcare: Planning and Infrastructure; Individual, Group, Organizational, and System Factors; The Process of Change; and Performance Measures and Evaluation. CONCLUSION: Our scan identified a tremendous amount of work being done around the world to improve healthcare. In general, our findings suggest these initiatives tend to be fragmented from an implementation standpoint. We identified primary and secondary drivers (factors) that can be used by those responsible for implementing large-scale improvement initiatives both at a strategy level and in their daily work. These drivers could serve as a "checklist" of ideas to consider in different testing and implementation situations.