The Healthcare Cost and Utilization Project (HCUP) comprises a series of national and state-level databases.³ This study used data from the Nationwide Inpatient Sample (NIS) and the Nationwide Emergency Department Sample (NEDS) database. This study was exempt from institutional review.

The NIS is created by drawing a stratified probability sample of hospitals drawn from State Inpatient Databases that include hospitalizations by patients with Medicare, Medicaid, private insurance, and the uninsured. From 2001 to 2011, the numbers of participating states grew from 33 to 46, the number of sampled hospitals ranged from 986 to 1,049, and the number of unweighted discharges ranged from 7,452,727 to 8,158,381. In 2012, 44 states, 4,378 hospitals, and 7,296,968 discharges were included covering about 95% of the US population. Prior to 2012, approximately 20% of all community hospitals as defined by the American Hospital Association were sampled from strata created on the basis of ownership/control, bed size, teaching status, urban/rural location, and US region, and all discharges from each hospital were used.⁴ In 2012, however, the NIS, renamed the National Inpatient Sample, selected a sample of 20% of all discharges from hospitals that were arranged in five strata.⁵ Hospital definitions previously based on American Hospital Association were changed to reflect those used by states. Sampling weights were created to produce national estimates. Because the NIS implemented a systematic sampling design in 2012, new sampling weights for trend analyses were developed.

Results for number, rate, length of stay, case-fatality rate, cost, and charge of hospitalizations for COPD and bronchiectasis were produced. In addition, the 30-day readmission rate for the years 2009 to 2012 were examined using three metrics: readmissions with the same International Classification of Diseases 9th Revision-Clinical Modification (ICD-9-CM) codes (see later discussion in this section) as the principal diagnosis, readmissions with the same ICD-9-CM codes as any diagnosis, and readmissions for any reason. Results were limited to nonneonatal, non-maternal discharges.

Costs were calculated from charges that were billed by the hospitals using cost-to-charge ratios produced by the Centers for Medicare & Medicaid Services from hospital accounting reports. When a cost-to-charge ratio was available for a specific hospital, that ratio was used to estimate the cost. For other hospitals, average cost-to-charge ratios for strata defined on the basis of bed size, teaching status, location, ownership, and region were used.

The NEDS database, which is the largest all-payer ED database in the United States, was used to examine trends in the numbers and rate of ED visits from 2006 to 2011.^6,7 NEDS is created by drawing stratified probability samples from community, nonrehabilitation hospitals that are part of the State Inpatient Databases and State Emergency Department Databases. A 20% sample is drawn from strata defined on the basis of region, trauma designation, urban-rural location, teaching status, and hospital ownership or control. The number of participating states ranged from 24 in 2006 to 30 in 2011, and the number of unweighted records ranged from 25.7 million in 2006 to 29.4 million in 2011. Sampling weights were created to produce national estimates.

The principal diagnosis listing ICD-9-CM codes that are part of the Clinical Classifications Software (CCS) diagnostic category 127 (COPD and bronchiectasis) was used to identify hospitalizations and ED visits. This category includes the following ICD-9-CM codes: 490, 491.0, 491.1, 491.2, 491.20, 491.21, 491.22, 491.8, 491.9, 492.0, 492.8, 494, 494.0, 494.1, and 496. By using the principal diagnosis, confidence that the hospitalization or ED visit was for COPD or bronchiectasis was maximized.

To compare rates of hospitalization and ED visit for COPD and bronchiectasis with several other conditions, rates for acute myocardial infarction (CCS category 100 comprising ICD-9-CM codes 410.0, 410.00, 410.01, 410.02, 410.1, 410.10, 410.11, 410.12, 410.2, 410.20, 410.21, 410.22, 410.3, 410.30, 410.31, 410.32, 410.4, 410.40, 410.41, 410.42, 410.5, 410.50, 410.51, 410.52, 410.6, 410.60, 410.61, 410.62, 410.7, 410.70, 410.71, 410.72, 410.8, 410.80, 410.81, 410.82, 410.9, 410.90, 410.91, 410.92), coronary atherosclerosis and other heart disease (CCS category 101 comprising ICD-9-CM codes 411.0, 411.1, 411.8, 411.81, 411.89, 412, 413.0, 413.1, 413.9, 414.0, 414.00, 414.01, 414.06, 414.2, 414.3, 414.4, 414.8, 414.9, V45.81, V45.82), nonhypertensive congestive heart failure (CCS category 108 comprising ICD-9-CM codes 398.91, 428.0, 428.1, 428.20, 428.21, 428.22, 428.23, 428.30, 428.31, 428.32, 428.33, 428.40, 428.41, 428.42, 428.43, 428.9), pneumonia (CCS category 122 comprising ICD-9-CM codes 003.22, 020.3, 020.4, 020.5, 021.2, 022.1, 031.0, 039.1, 052.1, 055.1, 073.0, 083.0, 112.4, 114.0, 114.4, 114.5, 115.05, 115.15, 115.95, 130.4, 136.3, 480.0, 480.1, 480.2, 480.3, 480.8, 480.9, 481, 482.0, 482.1, 482.2, 482.3, 482.30, 482.31, 482.32, 482.39, 482.4, 482.40, 482.41, 482.42, 482.49, 482.8, 482.81, 482.82, 482.83, 482.84, 482.89, 482.9, 483, 483.0, 483.1, 483.8, 484.1, 484.3, 484.5, 484.6, 484.7, 484.8, 485, 486, 513.0, 517.1), and lung cancer (CCS category 19 comprising ICD-9-CM codes 162.2, 162.3, 162.4, 162.5, 162.8, 162.9, 209.21, 231.2, V10.11) were calculated.

The online query system HCUPnet (http://hcupnet.ahrq.gov/) was used to generate data for this analysis.⁸ Analyses were limited to adults aged ≥ 18 years. Revised sampling weights that were produced as part of the 2012 redesign were used to generate national estimates for measures related to hospital discharges. Direct age adjustment to the projected year 2000 US population was performed using four age groups (18-44 years, 45-64 years, 65-84 years, and ≥ 85 years). Weighted least-squares regression was used to perform tests of linear trend for possible trends of the outcomes of interest. The online query system used the software SUDAAN (Research Triangle Institute) to generate SEs.

The weighted number of hospital discharges for COPD and bronchiectasis ranged from 534,398 to 692,364 (Fig 1). Of the numbers of combined COPD and bronchiectasis hospitalizations, 98.1% to 98.6% were for COPD (ICD-9 codes 490, 491, 492, 496). The average annual number of discharges was about 100,000 higher from 2008 to 2012 than from 2001 to 2007. The age-adjusted rate of discharges ranged from 242.7 to 286.0 per 100,000 population and showed no clear trend (P trend = .554) (Table 1). Significant decreases were noted for participants aged 18 to 44 years and 65 to 84 years, whereas a significant increase was noted for those aged 45 to 64 years. The mean length of stay decreased overall and in all age groups except in the youngest adults. The age-adjusted case-fatality rate dropped from 2.5% to 1.1%, and significant decreases occurred in all age groups. In contrast to COPD and bronchiectasis, the age-adjusted rates for discharges for acute myocardial infarction (P trend < .001), coronary atherosclerosis and other heart disease (P trend < .001), congestive heart failure (P trend < .001), pneumonia ( P trend < .001), and lung cancer (P trend < .001) declined significantly (Fig 2).

Aggregate charges for inpatient stays increased progressively from $8,023,983,422 in 2001 to $18,112,392,566 in 2012 (Fig 3) as did the mean and median charge per inpatient stay (Table 1). Mean and median costs increased as well but less sharply than charges.

The numbers of ED visits increased steadily from about 1,480,363 in 2006 to 1,787,612 in 2011 (Table 2). The age-adjusted rate increased mildly, driven by a significant increase in the rate among adults aged 45 to 64 years. Of those who had an ED visit, about 19% to 21% of these patients were admitted to the hospital. The age-specific mortality in the ED ranged from 0.1% to 0.3%.

Compared with COPD and bronchiectasis, the age-adjusted rate of ED visits declined for acute myocardial infarction (P trend = .002), coronary atherosclerosis and other heart disease (P trend < .001), congestive heart failure (P trend = .001), pneumonia (P trend = .044), and lung cancer (P trend < .001) (Fig 4).

Of patients with an index diagnosis of COPD or bronchiectasis, about 7% were readmitted within 30 days with COPD or bronchiectasis as the principal diagnosis, 18% were readmitted within 30 days with COPD or bronchiectasis listed as any diagnosis, and 21% were readmitted within 30 days for any reason (Table 3). The most recent estimates from 2012 suggested some improvement in readmission rates.

The current study, which provides the most up-to-date statistics, shows no significant trend in the age-adjusted rates of hospital discharges from 2001 to 2012 or ED visits from 2006 to 2011 for COPD or bronchiectasis. The persistence of these rates in the face of large declines in the prevalence of smoking is a bit perplexing but underscores the protracted nature of COPD. Significantly decreasing trends were found for mean length of stay and for the in-hospital case-fatality rate.

The lack of declining trend in the rate of hospital discharges in the NIS data contrasts with results from the NHDS that showed a decline in the age-adjusted rate from 402 per 100,000 in 1999 to 322 per 100,000 population in 2010. If 2001 is used as a starting point, however, the rate declined from 366 to 322 per 100,000 population in 2010.¹ It is noteworthy that the age range in the NHDS analysis was ≥ 25 years compared with ≥ 18 years in the present study. The lack of a significant trend in the age-adjusted rate of ED visits agrees with results from the NHAMCS.¹

The present study found a decline in mean length of stay from 4.5 days in 2001 to 4.0 days in 2012 among patients hospitalized for COPD or bronchiectasis. Limited information on temporal trends in length of stay among US patients with COPD has been published. In greater metropolitan Cleveland, Ohio, the mean length of stay for COPD hospitalizations decreased from 7.5 days in 1991 to 5.4 days in 1997.⁹ These declines could be due to improved outpatient and inpatient treatment, less severe forms of the disease, and pressures to reduce the costs of hospitalizations.

The NIS data showed a substantial decline in the in-hospital case-fatality rate from 2001 to 2012. A few reports have previously examined temporal changes in the in-hospital case-fatality rate in the United States. Data from the US Veterans Administration system showed that the in-hospital case-fatality rate among patients with COPD rose from 4.3% in 1970 to 7.9% in 1987.¹⁰ In greater metropolitan Cleveland, the unadjusted in-hospital mortality rate following a hospitalization for COPD among Medicare patients decreased by 49.6% from 1991 to 1997.⁹ Adjusted in-hospital and 30-day mortality rates decreased by 53.7% and 26%, respectively. The reasons for the large decline in the in-hospital case-fatality rate for hospitalizations for COPD or bronchiectasis could be due to improved treatment or to less severe forms of the disease. Some of the decline may also be attributable to the decrease in the mean length of stay.

Rates of hospitalizations and ED visits are a reflection of the numbers of adults with COPD or bronchiectasis and, once someone has the condition, utilization of health-care services. Therefore, declines in the rates can be accomplished by reducing the numbers of adults who have COPD, by reducing the need for hospitalizations once an adult has developed COPD, or a combination of the two. Favorable trends in the prevalence of smoking and in air pollution should eventually reduce the prevalence of COPD.^11,12 Although limited evidence points to a possible decline in the prevalence of COPD since 1999, the prevalence of self-reported COPD remains stubbornly high.¹ Among patients with COPD or obstructive lung function, the prevalence of smoking has remained very high.^13,14 Therefore, increased efforts to promote smoking cessation in these patients may reduce their hospitalizations and visits to the emergency room. In addition, annual vaccinations for influenza, vaccinations for pneumococcal disease, and optimal management of their condition including pulmonary rehabilitation should also reduce this need.

Reducing readmissions has become an important objective of medical care in recent times.¹⁵ Only 4 years of NIS data about the readmission rate in patients hospitalized for COPD or bronchiectasis were available, and these results showed mostly little change in the 30-day readmission rate during that time, with about 7% of patients with COPD or bronchiectasis being readmitted within 30 days with COPD or bronchiectasis as the principal diagnoses although the results for 2012 hinted at some improvement. Previous studies have reported 30-day read-mission rates among Medicare beneficiaries with COPD of 22.6% from 2003 to 2004 and of 7.1% for COPD as the principal diagnosis, 17.3% for any diagnosis of COPD, and 20.5% for any diagnosis in 2008.^16,17 A host of patient, clinical, provider, and system factors have been found to predict readmissions among patients with COPD.^18-28 Interventions to reduce readmissions have met with mixed success, but some that may reduce readmissions include continuity with patient's primary care provider or pulmonologist,²⁹ coordination of discharges,³⁰ and pulmonary rehabilitation.³¹ However, the time frame for examining the rate of readmissions following these interventions was as long as 1 year rather than 30 days.

Several limitations warrant consideration. Because data for the present study were obtained by using an online research tool, stratification by other factors such as demographic factors was not feasible. Second, severity of the condition that led a patient to access medical care was not available, and, therefore, the possible contribution of changes in disease severity to changes in length of stay and case-fatality rate could not be examined. Third, the validity of the ICD-9-CM codes used to identify hospitalizations and ED visits for COPD and bronchiectasis in the NIS and NEDS has not been specifically tested. Limited studies of hospitalizations in other settings have shown sensitivities of 55% to 65%, specificities of 67% to 90%, and a positive predictive value of 50.4%.^32,33 In addition, a study of ED records showed positive predictive values ranging from 60% to 100% depending on the ICD-9 code.³⁴

In conclusion, data from the NIS show that rates of hospital discharges from 2001 to 2012, ED visits from 2006 to 2011, and 30-day readmission rate from 2009 to 2012 among patients with COPD or bronchiectasis changed nonsignificantly. In contrast, significant reductions in length of stay and case-fatality rate occurred in the hospital. These results should serve to inform efforts to optimize medical care for patients with COPD or bronchiectasis in the United States.