Association of weekend continuity of care with hospital length of stay.

OBJECTIVE: The purpose of this study was to evaluate the association of physician continuity of care with length of stay, likelihood of weekend discharge, in-hospital mortality and 30-day readmission. DESIGN: A cohort study of hospitalized medical patients. The primary exposure was the weekend usual provider continuity (UPC) over the initial weekend of care. This metric was adapted from an outpatient continuity of care index. Regression models were developed to determine the association between UPC and outcomes. SETTING: An academic medical center. MAIN OUTCOME MEASURE: Length of stay which was calculated as the number of days from the first Saturday of the hospitalization to the day of discharge. RESULTS: Of the 3391 patients included in this study, the prevalence of low, moderate and high UPC for the initial weekend of hospitalization was 58.7, 22.3 and 19.1%, respectively. When compared with low continuity of care, both moderate and high continuity of care were associated with reduced length of stay, with adjusted rate ratios of 0.92 (95% CI 0.86-1.00) and 0.64 (95% CI 0.53-0.76), respectively. High continuity of care was associated with likelihood of weekend discharge (adjusted odds ratio 2.84, 95% CI 2.11-3.83) but was not significantly associated with mortality (adjusted odds ratio 0.72, 95% CI 0.29-1.80) or readmission (adjusted odds ratio 0.88, 95% CI 0.68-1.14) when compared with low continuity of care. CONCLUSIONS: Increased weekend continuity of care is associated with reduced length of stay. Improvement in weekend cross-coverage and patient handoffs may be useful to improve clinical outcomes.

Electronic health record use, intensity of hospital care, and patient outcomes.

OBJECTIVE: Previous studies have suggested that weekend hospital care is inferior to weekday care and that this difference may be related to diminished care intensity. The purpose of this study was to determine whether a metric for measuring intensity of hospital care based on use of the electronic health record was associated with patient-level outcomes. METHODS: We performed a cohort study of hospitalizations at an academic medical center. Intensity of care was defined as the hourly number of provider accessions of the electronic health record, termed "electronic health record interactions." Hospitalizations were categorized on the basis of the mean difference in electronic health record interactions between the first Friday and the first Saturday of hospitalization. We used regression models to determine the association of these categories with patient outcomes after adjusting for covariates. RESULTS: Electronic health record interactions decreased from Friday to Saturday in 77% of the 9051 hospitalizations included in the study. Compared with hospitalizations with no change in Friday to Saturday electronic health record interactions, the relative lengths of stay for hospitalizations with a small, moderate, and large decrease in electronic health record interactions were 1.05 (95% confidence interval [CI], 1.00-1.10), 1.11 (95% CI, 1.05-1.17), and 1.25 (95% CI, 1.15-1.35), respectively. Although a large decrease in electronic health record interactions was associated with in-hospital mortality, these findings were not significant after risk adjustment (odds ratio 1.74, 95% CI, 0.93-3.25). CONCLUSIONS: Intensity of inpatient care, measured by electronic health record interactions, significantly diminished from Friday to Saturday, and this decrease was associated with length of stay. Hospitals should consider monitoring and correcting temporal fluctuations in care intensity.

Monitoring the pulse of hospital activity: electronic health record utilization as a measure of care intensity.

BACKGROUND: Hospital care on weekends has been associated with reduced quality and poor clinical outcomes, suggesting that decreases in overall intensity of care may have important clinical effects. We describe a new measure of hospital intensity of care based on utilization of the electronic health record (EHR). METHODS: We measured global intensity of care at our academic medical center by monitoring the use of the EHR in 2011. Our primary measure, termed EHR interactions, was the number of accessions of a patient's electronic record by a clinician, adjusted for hospital census, per unit of time. Our secondary measure was percent of total available central processing unit (CPU) power used to access EHR servers at a given time. RESULTS: EHR interactions were lower on weekend days as compared to weekdays at every hour (P < 0.0001), and the daytime peak in intensity noted each weekday was blunted on weekends. The relative rate and 95% confidence interval (CI) of census-adjusted record accessions per patient on weekdays compared with weekends were: 1.76 (95% CI: 1.74-1.77), 1.52 (95% CI: 1.50-1.55), and 1.14 (95% CI: 1.12-1.17) for day, morning/evening, and night hours, respectively. Percent CPU usage correlated closely with EHR interactions (r = 0.90). CONCLUSIONS: EHR usage is a valid and easily reproducible measure of intensity of care in the hospital. Using this measure we identified large, hour-specific differences between weekend and weekday intensity. EHR interactions may serve as a useful measure for tracking and improving temporal variations in care that are common, and potentially deleterious, in hospital systems.

Guided ordering: clinician interactions with complex order-sets.

BACKGROUND: Electronic order-sets increasingly ask clinicians to answer questions or follow algorithms. Cooperation with such requests has not been studied. SETTING: Internal Medicine service of an academic medical center. OBJECTIVE: We studied the accuracy of clinician responses to questions embedded in electronic admission and discharge order-sets. Embedded questions asked whether any of three "core" diagnoses was present; a response was required to submit orders. Endorsement of any diagnosis made available best-practice ordering screens for that diagnosis. DESIGN: Three reviewers examined 180 electronic records (8% of discharges), drawn equally (for each core diagnosis) from possible combinations of Yes/No responses on admission and discharge. In addition to noting responses, we identified whether the core diagnosis was coded, determined from notes whether the admitting clinician believed that diagnosis present, and sought clinical evidence of disease on admission. We also surveyed participating clinicians anonymously about practices in answering embedded questions. MEASUREMENTS: We measured occurrence of six admission and five discharge scenarios relating medical record evidence of disease to clinician responses about its presence. RESULTS: The commonest discordant pattern between response and evidence was a negative response to disease presence on admission despite both early clinical evidence and documentation. Survey of study clinicians found that 75% endorsed some intentional inaccuracy; the commonest reason given was that questions were sometimes irrelevant to the clinical situation at the point asked. CONCLUSION: Through faults in order-set design, limitations of software, and/or because of an inherent tendency to resist directed behavior, clinicians may often ignore questions embedded in order-sets.

Risk-adjusted mortality: problems and possibilities.

The ratio of observed-to-expected deaths is considered a measure of hospital quality and for this reason will soon become a basis for payment. However, there are drivers of that metric more potent than quality: most important are medical documentation and patient acuity. If hositals underdocument and therefore do not capture the full "expected mortality" they may be tempted to lower their observed/expected ratio by reducing "observed mortality" through limiting access to the very ill. Underdocumentation occurs because hospitals do not recognize, and therefore cannot seek to confirm, specific comorbidities conferring high mortality risk. To help hospitals identify these comorbidities, this paper describes an easily implemented spread-sheet for evaluating comorbid conditions associated, in any particular hospital, with each discharge. This method identifies comorbidities that increase in frequency as mortality risk increases within each diagnostic grouping. The method is inductive and therefore independent of any particular risk-adjustment technique.

How do internal medicine residency programs evaluate their resident float experiences?

OBJECTIVES: We surveyed the nation's internal medicine residency training program directors to determine the range and frequency of existing methods by which float experiences are evaluated. METHODS: We sent questionnaires to the program directors of all 396 internal medicine residency training program sites in the country. Information requested included program characteristics, months devoted to float experiences in each year of training, and the location and purpose of the rotation. Program directors were also asked to choose among descriptors characterizing the evaluative process. RESULTS: There were 139 responding programs (39%), 134 with data that could be aggregated. Responding programs were similar to all programs nationally in the distributions of size and university sponsorship. Overall, 76% of programs employed a night float for any period of time, and 71% currently had one, on average for 6.7 years. Mean months of float experience during residency was 2.4 months, significantly longer in programs that were not university based. Float experiences were evaluated in 89% of those programs who employed them, with ten different methods reported. University-based programs were significantly less likely to use chart review as a method of evaluation, but no other differences in methodology were significant. CONCLUSIONS: Float rotations are common among internal medicine residency training programs. Evaluative methods vary, but one or more are applied in the vast majority of programs.

Effect of implementing pain management standards.

BACKGROUND: Historical undertreatment of pain among inpatients has resulted in a national requirement for pain practice standards. OBJECTIVE: We hypothesized that adoption/promulgation of practice standards in January 2003 at 1 suburban teaching hospital progressively increased compliance with those standards and decreased pain. DESIGN: We retrospectively reviewed medical records each month during 2003, when pain standards were adopted with repeated, institution-wide, and nursing-unit-based interventions. Also, we reviewed discharges during 1 month in adjacent years. PATIENTS: We identified adult patients from 20 medical and surgical All-Payer Refined Disease Related Groupings (APRDRGs) in which opiate charges were most common in 2003. Among these, we considered patients actually receiving opiates and randomly chose equal numbers of matching subjects in each month of 2003. Matching was for APRDRG and complexity group. We also matched January 2003 discharges with those from January 2001, 2002, and 2004. MEASUREMENTS: For each patient, we captured 3 variables measuring standards compliance: percentage pain observations reported numerically, number of observations, and median time to reassessment after opiates. We also captured 3 pain variables: median pain score, rate of improvement in pain score, and total opiates dispensed. RESULTS: There were 360 qualifying discharges in 2003, and 75 in the other years. Numeric observations increased 15%, number of assessments 36%, and reassessment time decreased 60%. All changes were significant but occurred before standards implementation. Among pain measures, only rate of pain improvement changed, worsening slightly but significantly (-0.02 to -0.005 U/h), also before standards. CONCLUSIONS: Implementation of pain practice standards affected neither practice nor pain.

Pounds of cure: chart weight as a measure of service intensity.

Intensity of hospital services is often estimated by length of stay (LOS). Increasing demands for documentation in the medical record suggested to us an alternate method: weighing the chart. In a retrospective study, we compared LOS and chart weight as predictors of actual hospital costs at a community teaching hospital. We reviewed a sample of 123 patients randomly chosen from the medical service and stratified by phase of the academic year. Both least-squares regression and a multiple sampling/validating technique were used to derive mean cost per ounce of chart and per day of stay. Costs estimated from weight were within 7% and from LOS within 14% of measured actual costs among patients not used to derive the formulae. We conclude that the intensity of paper documentation closely reflects actual costs; the same may be true of bytes or key-strokes for electronic records.

Assigning patients with heart failure to observation status: B-type natriuretic peptide, ejection fraction, or physician judgment.

It is of clinical and financial importance to identify those heart failure patients who are likely to improve rapidly. The authors evaluated, as predictors of short-term resolution, three clinical variables often used to predict long-term outcome. Consenting patients admitted to the emergency department with dyspnea were examined daily until resolution (symptom reversion to baseline absent worsening clinical signs or x-ray). The authors then compared hours to resolution of heart failure with serum B-type natriuretic peptide (BNP), ejection fraction (EF), and admitting physician prediction. They calculated negative and positive predictive values for EF, BNP, and physician opinion using quintile and published severity standards as cutoffs (for EF and BNP). Among 85 patients, BNP <400 pg/mL and <1000 pg/mL had poor positive predictive values (34% and 22%, respectively), as did EF (21%) and physician opinion (16%-21%, depending on physician confidence). Combining tests did not improve prediction. These results do not support the use of BNP, EF, or clinical assessment in triaging heart failure patients to short-stay status.

Assessing quality and efficiency of discharge summaries.

Discharge summaries are intended to transfer important clinical information from inpatient to outpatient settings and between hospital admissions. The authors created a point scale that rated summaries in 4 key areas and applied the scale at a community teaching hospital over 3 years. Charts of 150 patients were selected equally from those discharged early and late in the academic year. Residents dictated all summaries after July 2003 using a prominently displayed template. Two residents and a senior physician assessed dictation quality. Considerable differences were found among raters, particularly in the evaluation of style. The average of the 3 raters' scores improved 21%, and dictation length decreased 67% after introduction of the template (P < .001). No relationship was found among service intensity (measured as chart weight), dictation length (measured in lines), and quality. Measured by a comprehensive rating scale, the quality of discharge summaries increased with use of a template while their length decreased.

Cardiovascular complications after GI endoscopy: occurrence and risks in a large hospital system.

BACKGROUND: There is limited information concerning the risks for, and occurrence of, cardiovascular complications because of GI endoscopy. Published data are based on questionnaire surveys, which have a potential for bias. Moreover, available studies pertain exclusively to out-patients. METHODS: In-patients and day-stay patients who incurred charges for endoscopy with endoscopic procedure coding from 1999 through 2001 were identified from a financial database for all 9 hospitals in a large health care system. From these patients, those considered "at risk" for cardiovascular complications were selected based on charges for cardioactive medications, cardiac enzyme determinations, or intensive care services on the day of or the day after endoscopy. Medical records were reviewed for 25% of these patients, selected at random, noting demographics, history, and a modified Goldman score in patients with cardiovascular complications (defined as arrhythmia, chest pain or anginal equivalent, hypotension or myocardial infarction occurring within 24 hours after endoscopy). Identical information was obtained from a random sample of 0.5% of the chart records for all patients undergoing endoscopy. RESULTS: Patients who underwent endoscopy were not reliably identified for one hospital. This hospital was omitted from the calculation of the extrapolated rate of complication occurrence, but patients identified through chart review as having or not having a complication after endoscopy were included in the risk analysis. The extrapolated rate of occurrence of cardiovascular complications was 308: 95% CI [197, 457] per 100,000 procedures. Independent risk factors were: male gender, modified Goldman score, and use of propofol. CONCLUSIONS: In this study of patients undergoing hospital-based GI endoscopy, the risk of procedure-related cardiovascular complications was 2 to 70 times higher than previously reported. This finding may be ascribed to differences in the populations sampled and to a case-finding method that minimized reporting and ascertainment biases.

A randomized trial of initial warfarin dosing based on simple clinical criteria.

Warfarin induction is accomplished by titrating dosage to coagulation test results. Algorithms can guide this process but not identify the starting dose. We hypothesized that an initial warfarin dose approximating the maintenance value would safely enhance rapidity of induction. In a randomized trial we compared a fixed-dose to a maintenance-dose strategy for beginning warfarin therapy. To predict the maintenance dose among patients with differing warfarin requirements we performed regression analysis on clinical factors derived from chart review. Four community hospitals supplied records for retrospective analysis. The prospective trial was conducted in one, a 350-bed teaching institution. A sample of inpatients anticoagulated during 1998 formed the development set for retrospective study; a 1999 sample formed the validation set. A one year trial recruited consecutive eligible inpatients initiated on warfarin. We randomly assigned patients to a first warfarin dose calculated using our regression formula or fixed at 5 mg. All patients' subsequent doses were determined (as a percentage of initial) from coagulation testing. We compared days to anticoagulation, hospitalized hours, complications, and activity of factor II and protein C in a patient sample at intervals after induction. Weight, age, serum albumin, and presence of malignancy explained 25-30% of variance in maintenance dose. Ninety patients (44 calculated-dose and 46 standard-dose) evaluated in the clinical trial. Mean time to anticoagulation (among patients achieving anticoagulation) was 4.2 and 5.0 days, respectively (p = 0.007). We observed no significant differences in other endpoints. Individualized initial dosing may safely hasten warfarin induction.

Can computerized cost data substitute for chart review?

Identification of alternatives to manual chart review might improve efficiency in quality improvement work. This study at a large community teaching hospital in central New Jersey considered whether selected charges from a patient-level costs database could identify compliance with Sixth Scope of Work indicators in congestive heart failure (CHF). Charges resulting from specific tests, from test outcomes, and from prescribed treatments were identified from among 75 randomly chosen patients with CHF. In the sample 65% (as determined by database analysis) and 69% (as determined by chart review) complied with the principal peer review organization criterion. This difference was less than that found between review and re-review of study charts.

Actual and potential effects of medical resident coverage on reimbursement for inpatient visits by attending physicians.

CONTEXT: The impact of residents on hospital finance has been studied; there are no data describing the economic effect of residents on attending physicians. OBJECTIVE: In a community teaching hospital, we compared allowable inpatient visit codes and payments (based on documentation in the daily progress notes) between a general medicine teaching unit and nonteaching general medicine units. DESIGN: Retrospective chart review, matched cohort study. SETTING: Six hundred fifty-bed community teaching hospital. PATIENTS: Patients were discharged July 1998 through February 1999 from Saint Barnabas Medical Center. We randomly selected 200 patients in quartets. Each quartet consisted of a pair of patients cared for by residents and a pair cared for only by an attending physician. In each pair, 1 of the patients was under the care of an attending physician who usually admitted to the teaching service, and 1 was under the care of a usually nonteaching attending. Within each quartet, patients were matched for diagnosis-related group, length of stay, and discharge date. MAIN OUTCOME MEASURES: We assigned the highest daily visit code justifiable by resident and attending chart documentation, determining relative value units (RVUs) and reimbursements allowed by each patient's insurance company. RESULTS: Although more seriously ill, teaching-unit patients generated a mean 1.75 RVUs daily, compared with 1.84 among patients discharged from nonteaching units (P =.3). Median reimbursement, daily and per hospitalization, was similar on teaching and nonteaching units. Nonteaching attendings documented higher mean daily RVUs than teaching attendings (1.83 vs 1.76, P =.2). Median allowable reimbursements were $267 per case ($53 daily) among teaching attendings compared with $294 per case ($58 daily) among nonteaching attendings (Z = 1.54, P =.1). When only the resident note was considered, mean daily RVUs increased 39% and median allowable dollars per day 27% (Z = 4.21, P <.001). CONCLUSIONS: Nonteaching attendings appear to document their visits more carefully from a billing perspective than do teaching attendings. Properly counter-documented, resident notes could substantially increase payments to attending physicians.