Michael V. Murphy

The Preventability of Ventilator-Associated Events: The CDC Prevention Epicenters’ Wake Up and Breathe Collaborative

RATIONALE The CDC introduced ventilator-associated event (VAE) definitions in January 2013. Little is known about VAE prevention. We hypothesized that daily, coordinated spontaneous awakening trials (SATs) and spontaneous breathing trials (SBTs) might prevent VAEs. OBJECTIVES To assess the preventability of VAEs. METHODS We nested a multicenter quality improvement collaborative within a prospective study of VAE surveillance among 20 intensive care units between November 2011 and May 2013. Twelve units joined the collaborative and implemented an opt-out protocol for nurses and respiratory therapists to perform paired daily SATs and SBTs. The remaining eight units conducted surveillance alone. We measured temporal trends in VAEs using generalized mixed effects regression models adjusted for patient-level unit, age, sex, reason for intubation, Sequential Organ Failure Assessment score, and comorbidity index. MEASUREMENTS AND MAIN RESULTS We tracked 5,164 consecutive episodes of mechanical ventilation: 3,425 in collaborative units and 1,739 in surveillance-only units. Within collaborative units, significant increases in SATs, SBTs, and percentage of SBTs performed without sedation were mirrored by significant decreases in duration of mechanical ventilation and hospital length-of-stay. There was no change in VAE risk per ventilator day but significant decreases in VAE risk per episode of mechanical ventilation (odds ratio [OR], 0.63; 95% confidence interval [CI], 0.42-0.97) and infection-related ventilator-associated complications (OR, 0.35; 95% CI, 0.17-0.71) but not pneumonias (OR, 0.51; 95% CI, 0.19-1.3). Within surveillance-only units, there were no significant changes in SAT, SBT, or VAE rates. CONCLUSIONS Enhanced performance of paired, daily SATs and SBTs is associated with lower VAE rates. Clinical trial registered with www.clinicaltrials.gov (NCT 01583413).

Comparison of Trends in Sepsis Incidence and Coding Using Administrative Claims Versus Objective Clinical Data

BACKGROUND National reports of a dramatic rise in sepsis incidence are largely based on analyses of administrative databases. It is unclear if these estimates are biased by changes in coding practices over time. METHODS We calculated linear trends in the annual incidence of septicemia, sepsis, and severe sepsis at 2 academic hospitals from 2003 to 2012 using 5 different claims methods and compared case identification rates to selected objective clinical markers, including positive blood cultures, vasopressors, and/or lactic acid levels. RESULTS The annual incidence of hospitalizations with sepsis claims increased over the decade, ranging from a 54% increase for the method combining septicemia, bacteremia, and fungemia codes (P < .001 for linear trend) to a 706% increase for explicit severe sepsis/septic shock codes (P = .001). In contrast, the incidence of hospitalizations with positive blood cultures decreased by 17% (P = .006), and hospitalizations with positive blood cultures with concurrent vasopressors and/or lactic acidosis remained stable (P = .098). The sensitivity of sepsis claims for capturing hospitalizations with positive blood cultures with concurrent vasopressors and/or lactic acidosis increased (P < .001 for all methods), whereas the proportion of septicemia hospitalizations with positive blood cultures decreased from 50% to 30% (P < .001). CONCLUSIONS The incidence of hospitalizations with sepsis codes rose dramatically while hospitalizations with corresponding objective clinical markers remained stable or decreased. Coding for sepsis has become more inclusive, and septicemia diagnoses are increasingly being applied to patients without positive blood cultures. These changes likely explain some of the apparent rise in sepsis incidence and underscore the need for more reliable surveillance methods.

Descriptive Epidemiology and Attributable Morbidity of Ventilator-Associated Events

OBJECTIVE The Centers for Disease Control and Prevention implemented new surveillance definitions for ventilator-associated events (VAEs) in January 2013. We describe the epidemiology, attributable morbidity, and attributable mortality of VAEs. DESIGN Retrospective cohort study. SETTING Academic tertiary care center. PATIENTS All patients initiated on mechanical ventilation between January 1, 2006, and December 31, 2011. METHODS We calculated and compared VAE hazard ratios, antibiotic exposures, microbiology, attributable morbidity, and attributable mortality for all VAE tiers. RESULTS Among 20,356 episodes of mechanical ventilation, there were 1,141 (5.6%) ventilator-associated condition (VAC) events, 431 (2.1%) infection-related ventilator-associated complications (IVACs), 139 (0.7%) possible pneumonias, and 127 (0.6%) probable pneumonias. VAC hazard rates were highest in medical, surgical, and thoracic units and lowest in cardiac and neuroscience units. The median number of days to VAC onset was 6 (interquartile range, 4-11). The proportion of IVACs to VACs ranged from 29% in medical units to 42% in surgical units. Patients with probable pneumonia were more likely to be prescribed nafcillin, ceftazidime, and fluroquinolones compared with patients with possible pneumonia or IVAC-alone. The most frequently isolated organisms were Staphylococcus aureus (29%), Pseudomonas aeruginosa (14%), and Enterobacter species (7.9%). Compared with matched controls, VAEs were associated with more days to extubation (relative rate, 3.12 [95% confidence interval (CI), 2.96-3.29]), more days to hospital discharge (relative rate, 1.46 [95% CI, 1.37-1.55]), and higher hospital mortality risk (odds ratio, 1.98 [95% CI, 1.60-2.44]). CONCLUSIONS VAEs are common and morbid. Prevention strategies targeting VAEs are needed.

Risk factors for ventilator-associated events: a case-control multivariable analysis.

Objectives:The Centers for Disease Control and Prevention recently released new surveillance definitions for ventilator-associated events, including the new entities of ventilator-associated conditions and infection-related ventilator-associated complications. Both ventilator-associated conditions and infection-related ventilator-associated complications are associated with prolonged mechanical ventilation and hospital death, but little is known about their risk factors and how best to prevent them. We sought to identify risk factors for ventilator-associated conditions and infection-related ventilator-associated complications. Design:Retrospective case-control study. Setting:Medical, surgical, cardiac, and neuroscience units of a tertiary care teaching hospital. Patients:Hundred ten patients with ventilator-associated conditions matched to 110 controls without ventilator-associated conditions on the basis of age, sex, ICU type, comorbidities, and duration of mechanical ventilation prior to ventilator-associated conditions. Interventions:None. Measurements:We compared cases with controls with regard to demographics, comorbidities, ventilator bundle adherence rates, sedative exposures, routes of nutrition, blood products, fluid balance, and modes of ventilatory support. We repeated the analysis for the subset of patients with infection-related ventilator-associated complications and their controls. Main Results:Case and control patients were well matched on baseline characteristics. On multivariable logistic regression, significant risk factors for ventilator-associated conditions were mandatory modes of ventilation (odds ratio, 3.4; 95% CI, 1.6–8.0) and positive fluid balances (odds ratio, 1.2 per L positive; 95% CI, 1.0–1.4). Possible risk factors for infection-related ventilator-associated complications were starting benzodiazepines prior to intubation (odds ratio, 5.0; 95% CI, 1.3–29), total opioid exposures (odds ratio, 3.3 per 100 &mgr;g fentanyl equivalent/kg; 95% CI, 0.90–16), and paralytic medications (odds ratio, 2.3; 95% CI, 0.79–80). Traditional ventilator bundle elements, including semirecumbent positioning, oral care with chlorhexidine, venous thromboembolism prophylaxis, stress ulcer prophylaxis, daily spontaneous breathing trials, and sedative interruptions, were not associated with ventilator-associated conditions or infection-related ventilator-associated complications. Conclusions:Mandatory modes of ventilation and positive fluid balance are risk factors for ventilator-associated conditions. Benzodiazepines, opioids, and paralytic medications are possible risk factors for infection-related ventilator-associated complications. Prospective studies are needed to determine if targeting these risk factors can lower ventilator-associated condition and infection-related ventilator-associated complication rates.

What If Socioeconomics Made No Difference? Access To A Cadaver Kidney Transplant As An Example

OBJECTIVES Several studies have noted the impact of socioeconomic factors on access to expensive medical care, but none of those studies controlled for self-reported health and functional status or attitudes about treatment alternatives when analyses were completed. Because these factors may be correlated with socioeconomic status, the failure to control for them may have led to bias in other studies. The authors merged data from secondary sources with telephone survey data from a national sample of 456 end-stage renal disease patients to show how estimates of the effects of socioeconomic factors change when self-reported health and functional status and attitudes about treatment are incorporated into statistical models. The authors also showed how kidney transplant rates would change if socioeconomic factors no longer influences organ allocation decisions. METHODS Weibull proportional hazard analyses were used to show relationships between socioeconomic measures and waiting list entry and kidney transplant rates, before versus after accounting for self-reported health and functional status, attitudes about treatment, and other variables. Simulation analyses were used to estimate the number of waiting list spots and transplant operations that would move from economically advantaged to disadvantaged persons if socioeconomics no longer influenced organ allocation decisions. RESULTS Incorporating information about health and functional status, attitudes about treatment, and other factors into the hazard models often reduced the estimated impact of socioeconomic measures on the odds of (1) being on a waiting list for a cadaver kidney transplant and (2) receiving a transplant. Simulations showed that 30 to 65 waiting list spots or transplant operations per 1,000 patients would shift from economically advantaged to disadvantaged persons if socioeconomics no longer influenced organ allocation decisions. CONCLUSIONS Successful efforts to level the playing field would result in substantial redistributions of kidney transplants from economically advantaged to disadvantaged persons.

Improving documentation and coding for acute organ dysfunction biases estimates of changing sepsis severity and burden: a retrospective study

IntroductionClaims-based analyses report that the incidence of sepsis-associated organ dysfunction is increasing. We examined whether coding practices for acute organ dysfunction are changing over time and if so, whether this is biasing estimates of rising severe sepsis incidence and severity.MethodsWe assessed trends from 2005 to 2013 in the annual sensitivity and incidence of discharge ICD-9-CM codes for organ dysfunction (shock, respiratory failure, acute kidney failure, acidosis, hepatitis, coagulopathy, and thrombocytopenia) relative to standardized clinical criteria (use of vasopressors/inotropes, mechanical ventilation for ≥2 consecutive days, rise in baseline creatinine, low pH, elevated transaminases or bilirubin, abnormal international normalized ratio or low fibrinogen, and decline in platelets). We studied all adult patients with suspected infection (defined by ≥1 blood culture order) at two US academic hospitals.ResultsAcute organ dysfunction codes were present in 57,273 of 191,695 (29.9 %) hospitalizations with suspected infection, most commonly acute kidney failure (60.2 % of cases) and respiratory failure (28.9 %). The sensitivity of all organ dysfunction codes except thrombocytopenia increased significantly over time. This was most pronounced for acute kidney failure codes, which increased in sensitivity from 59.3 % in 2005 to 87.5 % in 2013 relative to a fixed definition for changes in creatinine (p = 0.019 for linear trend). Acute kidney failure codes were increasingly assigned to patients with smaller creatinine changes: the average peak creatinine change associated with a code was 1.99 mg/dL in 2005 versus 1.49 mg/dL in 2013 (p <0.001 for linear decline). The mean number of dysfunctional organs in patients with suspected infection increased from 0.32 to 0.59 using discharge codes versus 0.69 to 0.79 using clinical criteria (p <0.001 for both trends and comparison of the two trends). The annual incidence of hospitalizations with suspected infection and any dysfunctional organ rose an average of 5.9 % per year (95 % CI 4.3, 7.4 %) using discharge codes versus only 1.1 % (95 % CI 0.1, 2.0 %) using clinical criteria.ConclusionsCoding for acute organ dysfunction is becoming increasingly sensitive and the clinical threshold to code patients for certain kinds of organ dysfunction is decreasing. This accounts for much of the apparent rise in severe sepsis incidence and severity imputed from claims.

Objective Sepsis Surveillance Using Electronic Clinical Data

OBJECTIVE To compare the accuracy of surveillance of severe sepsis using electronic health record clinical data vs claims and to compare incidence and mortality trends using both methods. DESIGN We created an electronic health record-based surveillance definition for severe sepsis using clinical indicators of infection (blood culture and antibiotic orders) and concurrent organ dysfunction (vasopressors, mechanical ventilation, and/or abnormal laboratory values). We reviewed 1,000 randomly selected medical charts to characterize the definitions accuracy and stability over time compared with a claims-based definition requiring infection and organ dysfunction codes. We compared incidence and mortality trends from 2003-2012 using both methods. SETTING Two US academic hospitals. PATIENTS Adult inpatients. RESULTS The electronic health record-based clinical surveillance definition had stable and high sensitivity over time (77% in 2003-2009 vs 80% in 2012, P=.58) whereas the sensitivity of claims increased (52% in 2003-2009 vs 67% in 2012, P=.02). Positive predictive values for claims and clinical surveillance definitions were comparable (55% vs 53%, P=.65) and stable over time. From 2003 to 2012, severe sepsis incidence imputed from claims rose by 72% (95% CI, 57%-88%) and absolute mortality declined by 5.4% (95% CI, 4.6%-6.7%). In contrast, incidence using the clinical surveillance definition increased by 7.7% (95% CI, -1.1% to 17%) and mortality declined by 1.7% (95% CI, 1.1%-2.3%). CONCLUSIONS Sepsis surveillance using clinical data is more sensitive and more stable over time compared with claims and can be done electronically. This may enable more reliable estimates of sepsis burden and trends.

Lactate Testing in Suspected Sepsis: Trends and Predictors of Failure to Measure Levels.

Objectives:Serum lactate monitoring is central to risk stratification and management of sepsis and is now part of a potential quality measure. We examined 11-year trends in lactate testing and predictors of failure to measure lactates in patients with severe sepsis. Design:Retrospective cohort study. Setting:Two U.S. academic hospitals. Patients:Adult patients admitted from 2003 to 2013. Interventions:Annual rates of lactate measurement were assessed in patients who had blood cultures ordered and patients with severe sepsis, as defined by concomitant International Classification of Diseases, Ninth Revision codes for infection and organ dysfunction. The approximate time of suspected sepsis was determined by the first blood culture order with concurrent antibiotic initiation. Multivariate analysis was performed to identify predictors of failure to measure lactates in severe sepsis cases in 2013. Measurements and Main Results:Among hospitalizations with blood culture orders, rates of lactate measurement increased from 11% in 2003 to 48% in 2013 (p < 0.001 for linear trend). Rates of repeat lactate measurement within 6 hours after lactate levels greater than or equal to 4.0 mmol/L increased from 23% to 69% (p < 0.001). Patients were progressively less likely to be on vasopressors at the time of first lactate measurement (49% in 2003 vs 21% in 2013; p < 0.001). Despite these trends, lactates were measured at the time of suspected sepsis in only 65% of patients with severe sepsis in 2013. On multivariate analysis, hospital-onset sepsis and hospitalization on a nonmedical service were significant predictors of failure to measure lactates (adjusted odds ratio, 7.56; 95% CI, 6.31–9.06 and adjusted odds ratio, 2.08; 95% CI, 1.76–2.24, respectively). Conclusions:Lactate testing has increased dramatically over time and is being extended to patients without overt shock. However, rates of serial lactate testing are still suboptimal, and lactates are not being measured in many patients with severe sepsis. Hospital-onset sepsis and nonmedical units may be high-yield targets for quality improvement initiatives.

Improving Public Reporting and Data Validation for Complex Surgical Site Infections After Coronary Artery Bypass Graft Surgery and Hip Arthroplasty

Diagnosis codes in claims submitted for reimbursement following coronary artery bypass graft surgery and hip arthroplasty allow standardized and efficient identification of deep and organ/space surgical site infections.

Severity of Disease Estimation and Risk-Adjustment for Comparison of Outcomes in Mechanically Ventilated Patients Using Electronic Routine Care Data

BACKGROUND Valid comparison between hospitals for benchmarking or pay-for-performance incentives requires accurate correction for underlying disease severity (case-mix). However, existing models are either very simplistic or require extensive manual data collection. OBJECTIVE To develop a disease severity prediction model based solely on data routinely available in electronic health records for risk-adjustment in mechanically ventilated patients. DESIGN Retrospective cohort study. PARTICIPANTS Mechanically ventilated patients from a single tertiary medical center (2006-2012). METHODS Predictors were extracted from electronic data repositories (demographic characteristics, laboratory tests, medications, microbiology results, procedure codes, and comorbidities) and assessed for feasibility and generalizability of data collection. Models for in-hospital mortality of increasing complexity were built using logistic regression. Estimated disease severity from these models was linked to rates of ventilator-associated events. RESULTS A total of 20,028 patients were initiated on mechanical ventilation, of whom 3,027 deceased in hospital. For models of incremental complexity, area under the receiver operating characteristic curve ranged from 0.83 to 0.88. A simple model including demographic characteristics, type of intensive care unit, time to intubation, blood culture sampling, 8 common laboratory tests, and surgical status achieved an area under the receiver operating characteristic curve of 0.87 (95% CI, 0.86-0.88) with adequate calibration. The estimated disease severity was associated with occurrence of ventilator-associated events. CONCLUSIONS Accurate estimation of disease severity in ventilated patients using electronic, routine care data was feasible using simple models. These estimates may be useful for risk-adjustment in ventilated patients. Additional research is necessary to validate and refine these models.