Time-Series Approaches for Forecasting the Number of Hospital Daily Discharged Inpatients

For hospitals where decisions regarding acceptable rates of elective admissions are made in advance based on expected available bed capacity and emergency requests, accurate predictions of inpatient bed capacity are especially useful for capacity reservation purposes. As given, the remaining unoccupied beds at the end of each day, bed capacity of the next day can be obtained by examining the forecasts of the number of discharged patients during the next day. The features of fluctuations in daily discharges like trend, seasonal cycles, special-day effects, and autocorrelation complicate decision optimizing, while time-series models can capture these features well. This research compares three models: a model combining seasonal regression and ARIMA, a multiplicative seasonal ARIMA (MSARIMA) model, and a combinatorial model based on MSARIMA and weighted Markov Chain models in generating forecasts of daily discharges. The models are applied to three years of discharge data of an entire hospital. Several performance measures like the direction of the symmetry value, normalized mean squared error, and mean absolute percentage error are utilized to capture the under- and overprediction in model selection. The findings indicate that daily discharges can be forecast by using the proposed models. A number of important practical implications are discussed, such as the use of accurate forecasts in discharge planning, admission scheduling, and capacity reservation.

U.S. health care from a global perspective: spending, use of services, prices, and health in 13 countries

This analysis draws upon data from the Organization for Economic Cooperation and Development and other cross-national analyses to compare health care spending, supply, utilization, prices, and health outcomes across 13 high-income countries: Australia, Canada, Denmark, France, Germany, Japan, Netherlands, New Zealand, Norway, Sweden, Switzerland, the United Kingdom, and the United States. These data predate the major insurance provisions of the Affordable Care Act. In 2013, the U.S. spent far more on health care than these other countries. Higher spending appeared to be largely driven by greater use of medical technology and higher health care prices, rather than more frequent doctor visits or hospital admissions. In contrast, U.S. spending on social services made up a relatively small share of the economy relative to other countries. Despite spending more on health care, Americans had poor health outcomes, including shorter life expectancy and greater prevalence of chronic conditions.

[Hospital efficiency measured by bed space use in a secondary care hospital]

BACKGROUND

In recognition that the availability of resources in the medical facility forms part of the factors that influence the quality of healthcare, it is of vital importance to measure their outcome. The aim of this study was determine the efficiency of the medical facility through the use of beds in a secondary level hospital.

METHODS

Through the Health Information Management System (HIMS), we examined statistical reports from July 2012 to June 2013 including variables such as expenses, patient days, occupancy rate, average length of stay by specialty and medical division, results were obtained for each strategic indicator, and these results were related assumptions proposing to assess hospital efficiency.

RESULTS

Overall, we identified optimal efficiency of the medical facility without analysis of services, leads to deteriorating and low efficiency. The overall outcome of the five indicators applied overlooked saturation of services within the medical unit. However, the overall analysis shows the problem, noting the advantage of evaluating the same scenario from different perspectives.

CONCLUSIONS

The include indicators measuring hospital efficiency resource based bed, allows considering deficiencies identified, so that decision making is strengthened the decision making health.

Emergency Department Overcrowding and Ambulance Turnaround Time

OBJECTIVE

The aims of this study were to describe overcrowding in regional emergency departments in Seoul, Korea and evaluate the effect of crowdedness on ambulance turnaround time.

METHODS

This study was conducted between January 2010 and December 2010. Patients who were transported by 119-responding ambulances to 28 emergency centers within Seoul were eligible for enrollment. Overcrowding was defined as the average occupancy rate, which was equal to the average number of patients staying in an emergency department (ED) for 4 hours divided by the number of beds in the ED. After selecting groups for final analysis, multi-level regression modeling (MLM) was performed with random-effects for EDs, to evaluate associations between occupancy rate and turnaround time.

RESULTS

Between January 2010 and December 2010, 163,659 patients transported to 28 EDs were enrolled. The median occupancy rate was 0.42 (range: 0.10-1.94; interquartile range (IQR): 0.20-0.76). Overcrowded EDs were more likely to have older patients, those with normal mentality, and non-trauma patients. Overcrowded EDs were more likely to have longer turnaround intervals and traveling distances. The MLM analysis showed that an increase of 1% in occupancy rate was associated with 0.02-minute decrease in turnaround interval (95% CI: 0.01 to 0.03). In subgroup analyses limited to EDs with occupancy rates over 100%, we also observed a 0.03 minute decrease in turnaround interval per 1% increase in occupancy rate (95% CI: 0.01 to 0.05).

CONCLUSIONS

In this study, we found wide variation in emergency department crowding in a metropolitan Korean city. Our data indicate that ED overcrowding is negatively associated with turnaround interval with very small practical significance.

BED OCCUPANCY RATE AND LENGTH OF STAY OF PATIENTS IN MEDICAL AND ALLIED WARDS OF A TERTIARY CARE HOSPITAL

BACKGROUND

A good hospital management includes an effective allocative planning for beds in a hospital. Bed-occupancy rates and length of stay are the measures that reflect the functional ability of a hospital.

METHODS

A cross sectional study of two months' duration was carried out in eight medical and allied wards of Liaquat University Hospital (LUH) Jamshoro with objective to estimate the bed occupancy rate and the average length of stay of patients. Data was collected by filling a predesigned check list and bed occupancy rate and average length of stay were computed. Associations were analysed by using SPSS version 16. The p-value 50.05 Was taken as level of significance.

RESULTS

One hundred & seven admissions were recorded against 235 available beds. Average bed occupancy rate was 51.33%. The 51.4% of the patients in medical wards except paediatrics ward were of the age >50 years; the mean age was 45 years and standard deviation +/-6.4 years. Mean age in paediatrics was 3.89 years and standard deviation of +/-0.8 years. 55.1% patients had infectious diseases. The 32.7% patients stayed in hospital for up to 3 days showing significant association between nature of diseases and duration of stay (p=0.03). There was male preponderance, i.e., 54.2% males against 45.8% females. Showing significant association between gender and length of stay (p=0.01).

CONCLUSIONS

Bed occupancy rate and average stay in hospital were found within recommended range; more improvement may be brought by doing further research on this issue.

Nationwide cohort study of hip fractures: time trends in the incidence rates and projections up to 2035

SUMMARY

A growing elderly population is expected worldwide, and the burden of hip fractures on health care system will continue to increase. By 2035, there will be a 2.7-fold increase in the number of hip fractures in Taiwan. The study provides quantitative basis for the future distribution of medical resources.

INTRODUCTION

Hip fractures have long been recognized as a major public health concern. The study aimed to determine time trends in the incidence of hip fractures and to forecast the number of hip fractures expected in Taiwan up to 2035.

METHODS

A nationwide survey was conducted using data from the Taiwan National Health Insurance Research Database from 2004 to 2011. A total of 141,397 hip fractures were identified, with a mean of 17,675 fractures/year. Annual incidences of hip fractures were calculated and tested for trends. Projections of the incidence rates of hip fractures and bed days associated with hip fractures were calculated using Poisson regression on the historical incidence rates in combination with population projections from 2012 to 2035.

RESULTS

The incidence rates of hip fracture during 2004-2011 were 317 and 211 per 100,000 person-years among women and men, respectively. Over this 8-year period, the age-standardized incidence of hip fracture decreased by 13.4% among women and 12.2% among men. Despite the decline in the age-standardized incidence, the absolute number of hip fractures increased owing to the aging population. The number of hip fractures is expected to increase from 18,338 in 2010 to 50,421 in 2035-a 2.7-fold increase. The number of bed days for 2010 and 2035 was estimated at 161,248 and 501,995, respectively, representing a 3.1-fold increase.

CONCLUSIONS

The socioeconomic impact of hip fractures will be high in the near future. This study provides a quantitative basis for future policy decisions to serve this need.

Measuring the use of examination room time in oncology clinics: a novel approach to assessing clinic efficiency and patient flow

PURPOSE

The use of time in outpatient cancer clinics is a marker of quality and efficiency. Inefficiencies such as excessive patient wait times can have deleterious effects on clinic flow, functioning, and patient satisfaction. We propose a novel method of objectively measuring patient time in cancer clinic examination rooms and evaluating its impact on overall system efficiency.

METHODS

We video-recorded patient visits (N = 55) taken from a larger study to determine patient occupancy and flow in and out of examination rooms in a busy urban clinic in a National Cancer Institute-designated comprehensive cancer center. Coders observed video recordings and assessed patient occupancy time, patient wait time, and physician-patient interaction time. Patient occupancy time was compared with scheduled occupancy time to determine discrepancy in occupancy time. Descriptive and correlational analyses were conducted.

RESULTS

Mean patient occupancy time was 94.8 minutes (SD = 36.6), mean wait time was 34.9 minutes (SD = 28.8), and mean patient-physician interaction time was 29.0 minutes (SD = 13.5). Mean discrepancy in occupancy time was 40.3 minutes (range, 0.75 to 146.5 minutes). We found no correlation between scheduled occupancy time and patient occupancy time, patient-physician interaction time, and patient wait time, or between discrepancy in occupancy time and patient-physician interaction time.

CONCLUSION

The method is useful for assessing clinic efficiency and patient flow. There was no relationship between scheduled and actual time patients spend in exam rooms. Such data can be used in the design of interventions that reduce patient wait times, increase efficient use of resources, and improve scheduling patterns.

Probability of severe adverse events as a function of hospital occupancy

A unique application of regression modeling is described to compare hospital bed occupancy with reported severe adverse events amongst inpatients. The probabilities of the occurrence of adverse events as a function of hospital occupancy are calculated using logistic and multinomial regression models. All models indicate that higher occupancy rates lead to an increase in adverse events. The analysis identified that at an occupancy level of 100%, there is a 22% chance of one severe event occurring and a 28% chance of at least one severe event occurring. This modeling contributes evidence toward the management of hospital occupancy to benefit patient outcomes.

Estimation of hand hygiene opportunities on an adult medical ward using 24-hour camera surveillance: validation of the HOW2 Benchmark Study

BACKGROUND

We previously published a formula to estimate the number of hand hygiene opportunities (HHOs) per patient-day using the World Health Organization's "Five Moments for Hand Hygiene" methodology (HOW2 Benchmark Study). HHOs can be used as a denominator for calculating hand hygiene compliance rates when product utilization data are available. This study validates the previously derived HHO estimate using 24-hour video surveillance of health care worker hand hygiene activity.

METHODS

The validation study utilized 24-hour video surveillance recordings of 26 patients' hospital stays to measure the actual number of HHOs per patient-day on a medicine ward in a large teaching hospital. Statistical methods were used to compare these results to those obtained by episodic observation of patient activity in the original derivation study.

RESULTS

Total hours of data collection were 81.3 and 1,510.8, resulting in 1,740 and 4,522 HHOs in the derivation and validation studies, respectively. Comparisons of the mean and median HHOs per 24-hour period did not differ significantly. HHOs were 71.6 (95% confidence interval: 64.9-78.3) and 73.9 (95% confidence interval: 69.1-84.1), respectively.

CONCLUSION

This study validates the HOW2 Benchmark Study and confirms that expected numbers of HHOs can be estimated from the unit's patient census and patient-to-nurse ratio. These data can be used as denominators in calculations of hand hygiene compliance rates from electronic monitoring using the "Five Moments for Hand Hygiene" methodology.

The rising rate of admissions for tonsillitis and neck space abscesses in England, 1991-2011

INTRODUCTION

Sore throats and tonsillitis represent a considerable health burden as well as a significant source of expenditure for the National Health Service (NHS). As part of the recent NHS savings drive, the introduction of 'procedures of low clinical effectiveness' (PoLCE) lists has reinforced a large reduction in the number of tonsillectomies performed. We carried out a cross-sectional study of trends in emergency sore throat admissions in the context of the number of tonsillectomies performed.

METHODS

Hospital Episode Statistics (HES) data were extracted. Office for National Statistics data were also used.

RESULTS

Between 1991 and 2011, the overall tonsillectomy rate fell by 44%. In the same time, the admission rate for tonsillitis rose by 310% (Pearson's r=-0.67, p=0.01). The peritonsillar abscess admission rate rose by 31% (r=-0.79, p<0.01). Between 1996 and 2011, the overall tonsillectomy rate fell by 41% and the retro and parapharyngeal abscess admission rate rose by 39% (r=-0.55, p=0.026). There was a 14% overall increase in tonsillectomy and sore throat associated bed days. This was despite the large fall in tonsillectomy numbers and the reduction in length of hospital stay.

CONCLUSIONS

Efforts to reduce the tonsillectomy rate are correlated with a significant rise in emergency admissions. The rise in the retro and parapharyngeal abscess rate is perhaps most alarming given the very high mortality of these conditions. Bed day data suggest that no net saving has been made despite the new measures.

Hospital factors impact variation in emergency department length of stay more than physician factors

INTRODUCTION

To analyze the correlation between the many different emergency department (ED) treatment metric intervals and determine if the metrics directly impacted by the physician correlate to the "door to room" interval in an ED (interval determined by ED bed availability). Our null hypothesis was that the cause of the variation in delay to receiving a room was multifactorial and does not correlate to any one metric interval.

METHODS

We collected daily interval averages from the ED information system, Meditech©. Patient flow metrics were collected on a 24-hour basis. We analyzed the relationship between the time intervals that make up an ED visit and the "arrival to room" interval using simple correlation (Pearson Correlation coefficients). Summary statistics of industry standard metrics were also done by dividing the intervals into 2 groups, based on the average ED length of stay (LOS) from the National Hospital Ambulatory Medical Care Survey: 2008 Emergency Department Summary.

RESULTS

Simple correlation analysis showed that the doctor-to-discharge time interval had no correlation to the interval of "door to room (waiting room time)", correlation coefficient (CC) (CC=0.000, p=0.96). "Room to doctor" had a low correlation to "door to room" CC=0.143, while "decision to admitted patients departing the ED time" had a moderate correlation of 0.29 (p <0.001). "New arrivals" (daily patient census) had a strong correlation to longer "door to room" times, 0.657, p<0.001. The "door to discharge" times had a very strong correlation CC=0.804 (p<0.001), to the extended "door to room" time.

CONCLUSION

Physician-dependent intervals had minimal correlation to the variation in arrival to room time. The "door to room" interval was a significant component to the variation in "door to discharge" i.e. LOS. The hospital-influenced "admit decision to hospital bed" i.e. hospital inpatient capacity, interval had a correlation to delayed "door to room" time. The other major factor affecting department bed availability was the "total patients per day." The correlation to the increasing "door to room" time also reflects the effect of availability of ED resources (beds) on the patient evaluation time. The time that it took for a patient to receive a room appeared more dependent on the system resources, for example, beds in the ED, as well as in the hospital, than on the physician.

[Early team assessment relieves the emergency department]

A model for physician-led team triage was evaluated at the Emergency Department at the University hospital of Örebro, Sweden. Data from 1600 patients indicate that this work model reduces length of stay, time to physician assessment, emergency department occupancy, rate of admission and the proportion of patients in need of close monitoring. The project was conducted without any change in the number of physicians, nurses or staff nurses working in the Emergency Department. 

[The structural functional analysis of beds stock of curative preventive organizations of the state public health system of the Russian Federation]

The article presents the results of comprehensive scientific analysis of size and structure of beds stock of medical curative preventive organizations of state and municipal health care systems of the Russian Federation. The issues of beds support of population on national, federal okrugs and federation subjects' levels including differentiation on different medical specialties are considered. The main indicators of functioning of hospitals, per capita consumption of hospital medical care and territorial characteristics and differences of these indicators are analyzed In conditions of on-going decrease of size of beds stock and amount of medical care in hospitals and against the background of stability of main indicators of beds use the expressed but not always objectively conditioned differences continue to be present concerning both population support with beds stock and indicators of consumption of medical care in hospitals. All these occurrences undoubtedly impact accessibility of this type of medical care to population and its resource capacity for the government. In 2012, beds support of population decreased from 85.7 to 84.1 beds per 10 000 of population. The value of indicator in federal subjects differs up to 2.9 times. In the structure of beds stock are prevailing specialized beds or groups of beds on such medical specialties as psychiatry, surgery, obstetrics and gynecology and therapy. The per capita use of medical care in hospitals decreased up to 2.609 beds-per-day that is 6.2% lower than standard value from the program of state guarantees of free-of-charge medical care support of citizen. The end values of indicator in federal subjects differ in 2.7 times. In federal subjects indicators of mean number of work of bed per year differ up to 1.2 times, of mean duration of treatment--up to 1.6 times, turn-over of bed--up to 1.6 times, hospital lethality--up to 5.9 times. The results of study confirm necessity of structural functional optimization of national beds stock and development of interaction between hospitals and out-patient services.

Using geographic information systems to simulate patient access areas

We constructed a simulation model with a geographic information system (GIS) to predict the future shortage of beds in the Tokyo Metropolitan Area. With a grid square method, we calculated patient numbers for every 500 square meters of the Tokyo Metropolitan Area until 2040 and estimated whether those in need could be admitted to hospitals within an hour's drive from their homes. The simulation demonstrates that after 2025 many patients may not be able to find hospitals within this time framework. The situation will be especially serious in the center of Tokyo and along the railway lines, where many senior citizens reside. We can now apply this innovative GIS method in many fields and especially for the precise estimation of future demands for and supply of medical assistance.

Mid-term NEAT review: analysing the improvements in hospital ED performance

Introduced with a promise to reduce overcrowding in the Emergency Department (ED) and the associated morbidity and mortality linked to bed access difficulties, the National Emergency Access Target (NEAT) is now over halfway through transitionary arrangements towards a target of 90% of patients that visit a hospital ED being admitted or discharged within 4 hours. Facilitation and reward funding has ensured hospitals around the country are remodelling workflows to ensure compliance. Recent reports however show that the majority of hospitals are still far from being able to meet this target. We investigate the NEAT journey of 30 Queensland hospitals over the past two years and compare this performance to a previous study that investigated the 4 hour ED discharge performance of these hospitals at various times of day and under varying occupancy conditions. Our findings reveal that, while most hospitals have made significant improvements to their 4 hour discharge performance in 2013, the underlying flow patterns and periods of poor NEAT compliance remain largely unchanged. The work identifies areas for targeted improvement to inform system redesign and workflow planning.

The effect of a large Legionnaires' disease outbreak in Southwest Edinburgh on acute and critical care services

OBJECTIVE

The largest outbreak of Legionnaires Disease (LD) in the UK for a decade occurred in Edinburgh in June 2012. We describe the clinical and public health management of the outbreak.

SETTING

Three acute hospitals covering an urban area of ~480,000.

METHODS

Data were collected on confirmed and suspected cases and minutes of the Incident Management Team meetings were reviewed to identify key actions.

RESULTS

Over 1600 urine samples and over 600 sputum samples were tested during the outbreak. 61 patients with pneumonia tested positive for Legionella pneumophila serogroup 1 by urinary antigen detection, culture, respiratory PCR or serology. A further 23 patients with pneumonia were treated as suspected cases on clinical and epidemiological grounds but had no microbiological diagnosis. 36% of confirmed and probable cases required critical care admission. Mean ICU length of stay was 11.3 (±7.6) days and mean hospital length of stay for those who were admitted to ICU was 23.0 (±17.2) days. For all hospitalized patients the mean length of stay was 15.7 (±14) days. In total there were four deaths associated with this outbreak giving an overall case fatality of 6.5%. Hospital and critical care mortality was 6.1% and 9.1%, respectively.

CONCLUSION

A significant proportion of patients required prolonged multiple organ support or complex ventilation. Case fatality compared favourably to other recent outbreaks in Europe. Access to rapid diagnostic tests and prompt antibiotic therapy may have mitigated the impact of pre-existing poor health among those affected.

ICU occupancy and mechanical ventilator use in the United States

OBJECTIVES

Detailed data on occupancy and use of mechanical ventilators in U. S. ICU over time and across unit types are lacking. We sought to describe the hourly bed occupancy and use of ventilators in U.S. ICUs to improve future planning of both the routine and disaster provision of intensive care.

DESIGN

Retrospective cohort study. We calculated mean hourly bed occupancy in each ICU and hourly bed occupancy for patients on mechanical ventilators. We assessed trends in overall occupancy over the 3 years. We also assessed occupancy and mechanical ventilation rates across different types and sizes of ICUs.

SETTING

Ninety-seven U.S. ICUs participating in Project IMPACT from 2005 to 2007.

PATIENTS

A total of 226,942 consecutive admissions to ICUs.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

Over the 3 years studied, total ICU occupancy ranged from 57.4% to 82.1% and the number of beds filled with mechanically ventilated patients ranged from 20.7% to 38.9%. There was no change in occupancy across years and no increase in occupancy during influenza seasons. Mean hourly occupancy across ICUs was 68.2% ± 21.3% (SD) and was substantially higher in ICUs with fewer beds (mean, 75.8% ± 16.5% for 5-14 beds vs 60.9% ± 22.1% for 20+ beds, p = 0.001) and in academic hospitals (78.7% ± 15.9% vs 65.3% ± 21.3% for community not-for-profit hospitals, p < 0.001). More than half of ICUs (53.6%) had 4+ beds available more than half the time. The mean percentage of ICU patients receiving mechanical ventilation in any given hour was 39.5% (± 15.2%), and a mean of 29.0% (± 15.9%) of ICU beds were filled with a patient on a ventilator.

CONCLUSIONS

Occupancy of U.S. ICUs was stable over time, but there is uneven distribution across different types and sizes of units. Only three of 10 beds were filled at any time with mechanically ventilated patients, suggesting substantial surge capacity throughout the system to care for acutely critically ill patients.

Effects of multidisciplinary teamwork on lead times and patient flow in the emergency department: a longitudinal interventional cohort study

BACKGROUND

Long waiting times for emergency care are claimed to be caused by overcrowded emergency departments and non-effective working routines. Teamwork has been suggested as a promising solution to these issues. The aim of the present study was to investigate the effects of teamwork in a Swedish emergency department on lead times and patient flow.

METHODS

The study was set in an emergency department of a university hospital where teamwork, a multi-professional team responsible for the whole care process for a group of patients, was introduced. The study has a longitudinal non-randomized intervention study design. Data were collected for five two-week periods during a period of 1.5 years. The first part of the data collection used an ABAB design whereby standard procedure (A) was altered weekly with teamwork (B). Then, three follow-ups were conducted. At last follow-up, teamwork was permanently implemented. The outcome measures were: number of patients handled within teamwork time, time to physician, total visit time and number of patients handled within the 4-hour target.

RESULTS

A total of 1,838 patient visits were studied. The effect on lead times was only evident at the last follow-up. Findings showed that the number of patients handled within teamwork time was almost equal between the different study periods. At the last follow-up, the median time to physician was significantly decreased by 11 minutes (p = 0.0005) compared to the control phase and the total visit time was significantly shorter at last follow-up compared to control phase (p = <0.0001; 39 minutes shorter on average). Finally, the 4-hour target was met in 71% in the last follow-up compared to 59% in the control phase (p = 0.0005).

CONCLUSIONS

Teamwork seems to contribute to the quality improvement of emergency care in terms of small but significant decreases in lead times. However, although efficient work processes such as teamwork are necessary to ensure safe patient care, it is likely not sufficient for bringing about larger decreases in lead times or for meeting the 4-hour target in the emergency department.