A Risk‐Adjusted O–E CUSUM with Monitoring Bands for Monitoring Medical Outcomes

Sun, Rena Jie; Kalbfleisch, John D.

doi:10.1111/j.1541-0420.2012.01822.x

Cited by 16 publications

(26 citation statements)

References 17 publications

(41 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Hospital mortality monitoring studies were international, covering the UK, 2,14,35,51,53 the USA, 49,52,59,64,67,68,70 the Netherlands, 57,63 Canada 55,69 and Australia and New Zealand. 50,54,56,58,[60][61][62]65,66 Studies were for a range of conditions and settings; these included surgical, 2,51,59,64,66,70 intensive care, 50,52,54,61,63,65 transplantation 53,67,68 and all cause. 14,49,55,57,58,60 …”

Section: Overview Of Approachmentioning

confidence: 99%

Evaluation of a national surveillance system for mortality alerts: a mixed-methods study

Aylin

Bottle

Burnett

et al. 2018

Health Serv Deliv Res

View full text Add to dashboard Cite

BackgroundSince 2007, Imperial College London has generated monthly mortality alerts, based on statistical process control charts and using routinely collected hospital administrative data, for all English acute NHS hospital trusts. The impact of this system has not yet been studied.ObjectivesTo improve understanding of mortality alerts and evaluate their impact as an intervention to reduce mortality.DesignMixed methods.SettingEnglish NHS acute hospital trusts.ParticipantsEleven trusts were included in the case study. The survey involved 78 alerting trusts.Main outcome measuresRelative risk of mortality and perceived efficacy of the alerting system.Data sourcesHospital Episodes Statistics, published indicators on quality and safety, Care Quality Commission (CQC) reports, interviews and documentary evidence from case studies, and a national evaluative survey.MethodsDescriptive analysis of alerts; association with other measures of quality; associated change in mortality using an interrupted time series approach; in-depth qualitative case studies of institutional response to alerts; and a national cross-sectional evaluative survey administered to describe the organisational structure for mortality governance and perceptions of efficacy of alerts.ResultsA total of 690 mortality alerts generated between April 2007 and December 2014. CQC pursued 75% (154/206) of alerts sent between 2011 and 2013. Patient care was cited as a factor in 70% of all investigations and in 89% of sepsis alerts. Alerts were associated with indicators on bed occupancy, hospital mortality, staffing, financial status, and patient and trainee satisfaction. On average, the risk of death fell by 58% during the 9-month lag following an alert, levelling afterwards and reaching an expected risk within 18 months of the alert. Acute myocardial infarction (AMI) and sepsis alerts instigated institutional responses across all the case study sites, although most sites were undertaking some parallel activities at a more general level to address known problems in care in these and other areas. Responses included case note review and coding improvements, changes in patient pathways, changes in diagnosis of sepsis and AMI, staff training in case note write-up and coding, greater transparency in patient deterioration, and infrastructure changes. Survey data revealed that 86% of responding trusts had a dedicated trust-level lead for mortality reduction and 92% had a dedicated trust-level mortality group or committee in place. Trusts reported that mortality reduction was a high priority and that there was strong senior leadership support for mortality monitoring. The weakest areas reported concerned the accuracy of coding, the quality of specialty-level mortality data and understanding trends in specialty-level mortality data.LimitationsOwing to the correlational nature of our analysis, we could not ascribe a causal link between mortality alerts and reductions in mortality. The complexity of the institutional context and behaviour hindered our capacity to attribute locally reported changes specifically to the effects of the alerts rather than to ongoing institutional strategy.ConclusionsThe mortality alert surveillance system reflects aspects of quality care and is valued by trusts. Alerts were considered a useful focus for identifying problems and implementing interventions around mortality.Future workA further analysis of site visits and survey material, the application of evaluative framework to other interventions, a blinded case note review and the dissemination of findings.FundingThe National Institute for Health Research Health Services and Delivery Research programme.

show abstract

Section: Overview Of Approachmentioning

confidence: 99%

Evaluation of a national surveillance system for mortality alerts: a mixed-methods study

Aylin

Bottle

Burnett

et al. 2018

Health Serv Deliv Res

View full text Add to dashboard Cite

show abstract

“…When this plot trends upwards (downwards), the observed failure rates in this center are higher (lower) than those in the reference population, and so these plots provide a useful descriptive tool. Further discussion can be found in [7] or [8]. See also [9].…”

Section: Definition and Properties Of The Weighted O-e Cusummentioning

confidence: 99%

“…Refer to Figure 1 for an example chart designed to detect worse than expected signals, although this chart is revised to accommodate dependent censoring. Note that a one-sided CUSUM can be designed to help detect either a worse than expected performance with Â > 0 or a better than expected performance with Â < 0; see [7,10].…”

Section: The One-sided Cusum Chartmentioning

confidence: 99%

A weighted cumulative sum (WCUSUM) to monitor medical outcomes with dependent censoring

Sun

Kalbfleisch

Schaubel

2014

Statistics in Medicine

Self Cite

View full text Add to dashboard Cite

We develop a Weighted CUmulative SUM (WCUSUM) to evaluate and monitor pre-transplant waitlist mortality of facilities in the context where transplantation is considered to be dependent censoring. Waitlist patients are evaluated multiple times in order to update their current medical condition as reflected in a time dependent variable called the Model for End-Stage Liver Disease (MELD) score. Higher MELD scores are indicative of higher pre-transplant death risk. Moreover, under the current liver allocation system, patients with higher MELD scores receive higher priority for liver transplantation. To evaluate the waitlist mortality of transplant centers, it is important to take this dependent censoring into consideration. We assume a ‘standard’ transplant practice through a transplant model and utilize Inverse Probability Censoring Weights (IPCW) to construct a weighted CUSUM. We evaluate the properties of a weighted zero-mean process as the basis of the proposed weighted CUSUM. We then discuss a resampling technique to obtain control limits. The proposed WCUSUM is illustrated through the analysis of national transplant registry data.

show abstract

“…Profiling methods exist mostly for static or non time-varying patient outcomes, such as an annual risk-standardized mortality or annual 30-day hospital readmission rate (Normand et al 1997;Normand and Shahian 2007;Ash et al 2011;Horwitz et al 2011;He et al 2013;Kalbfleisch and Wolfe 2013;CMS 2014;and references therein). The limited timevarying metrics in the literature include the risk-adjusted CUmulative SUMmation (CUSUM) and observed-expected CUSUM techniques (Biswas and Kalbfleisch 2008;Sun and Kalbfleisch 2013;Van Rompaye et al 2015), although these are tailored to survival time outcomes. In profiling, the modeling process to evaluate a provider includes an important risk-adjustment step to account for patient case-mix (i.e., baseline patient-level comorbidities).…”

Section: Introductionmentioning

confidence: 99%

Time-Dynamic Profiling with Application to Hospital Readmission Among Patients on Dialysis

Estes

Nguyen

Chen³

et al. 2018

Biometrics

View full text Add to dashboard Cite

Standard profiling analysis aims to evaluate medical providers, such as hospitals, nursing homes, or dialysis facilities, with respect to a patient outcome. The outcome, for instance, may be mortality, medical complications, or 30-day (unplanned) hospital readmission. Profiling analysis involves regression modeling of a patient outcome, adjusting for patient health status at baseline, and comparing each provider's outcome rate (e.g., 30-day readmission rate) to a normative standard (e.g., national "average"). Profiling methods exist mostly for non time-varying patient outcomes. However, for patients on dialysis, a unique population which requires continuous medical care, methodologies to monitor patient outcomes continuously over time are particularly relevant. Thus, we introduce a novel time-dynamic profiling (TDP) approach to assess the time-varying 30-day readmission rate. TDP is used to estimate, for the first time, the risk-standardized time-dynamic 30-day hospital readmission rate, throughout the time period that patients are on dialysis. We develop the framework for TDP by introducing the standardized dynamic readmission ratio as a function of time and a multilevel varying coefficient model with facility-specific time-varying effects. We propose estimation and inference procedures tailored to the problem of TDP and to overcome the challenge of high-dimensional parameters when examining thousands of dialysis facilities.

show abstract

A Risk‐Adjusted O–E CUSUM with Monitoring Bands for Monitoring Medical Outcomes

Cited by 16 publications

References 17 publications

Evaluation of a national surveillance system for mortality alerts: a mixed-methods study

Evaluation of a national surveillance system for mortality alerts: a mixed-methods study

A weighted cumulative sum (WCUSUM) to monitor medical outcomes with dependent censoring

Time-Dynamic Profiling with Application to Hospital Readmission Among Patients on Dialysis

Contact Info

Product

Resources

About