Notable declines in both pneumonia hospitalizations and inpatient mortality attributed to pneumonia actually appear to be statistical artifacts related to changes in diagnostic coding rather than to bonafide improvements in health care, according to a report in the April 4 issue of JAMA.
A nationwide 27% reduction in pneumonia hospitalizations and an accompanying 28% reduction in pneumonia mortality were offset by concomitant rises in the rates of hospitalization and death due to sepsis (with a secondary diagnosis of pneumonia) and to respiratory failure (with a secondary diagnosis of pneumonia), said Dr. Peter K. Lindenauer of the Center for Quality of Care Research, Baystate Medical Center, Springfield, Mass., and his associates.
"These results suggest that secular trends in documentation and coding, rather than improvements in actual outcomes, may explain much of the observed change in this and other studies," they noted.
The findings also suggest that ratings of hospital performance based on pneumonia statistics may be inaccurate because of variation across hospitals in the use of diagnostic codes for pneumonia, sepsis, and respiratory failure, they added.
Noting that several epidemiologic studies have reported improvements in pneumonia statistics but that there haven’t been any "care-transforming technologies" to account for this improvement, Dr. Lindenauer and his colleagues analyzed trends in pneumonia hospital admissions and outcomes over time. They used data from the 2003 through the 2009 Nationwide Inpatient Sample (NIS), the largest all-payer hospital database in the country, which covers between 5 million and 8 million discharges each year. The NIS is sponsored by the U.S. Agency for Healthcare Research and Quality.
The researchers assessed hospitalizations for a principal diagnosis of pneumonia, as well as for a principal diagnosis of sepsis or respiratory failure together with a secondary diagnosis of pneumonia. For control conditions, they assessed hospitalizations for a principal diagnosis of ischemic stroke, ST-segment-elevation myocardial infarction (STEMI), and ruptured thoracic or abdominal aortic aneurysms.
"We also considered change over time in discharge disposition, including discharge to hospice, as a secondary outcome because increasing referral to inpatient nursing and rehabilitation facilities and hospice might allow sicker patients to be discharged rather than retained in the hospital," they noted.
From 2003 to 2009, the hospitalization rate of patients with a principal diagnosis of pneumonia decreased by 27.4%, from 5.5 per 1,000 to 4.0 per 1,000. This reversed "a well-documented decades-long trend toward increasing hospitalization" for the disorder, the investigators said (JAMA 2012;307:1405-13).
During the same period, however, the hospitalization rate for patients with a principal diagnosis of sepsis and a secondary diagnosis of pneumonia rose 177.6%, from 0.4 per 1,000 to 1.1 per 1,000. And the hospitalization rate for patients with a principal diagnosis of respiratory failure and a secondary diagnosis of pneumonia rose 9.3%, from 0.44 per 1,000 to 0.48 per 1,000.
These trends were consistent across all age groups and for both men and women.
During the same period, inpatient pneumonia mortality declined from 5.8% to 4.2%, a relative risk reduction (RRR) of 28.2%. There was a concomitant decline in inpatient sepsis mortality (RRR, 12%) and in inpatient respiratory failure mortality (RRR, 23.7%).
However, "when the three groups were combined ... there was little change in the inpatient mortality rate, varying from a small increase to a small decline, depending on the approach to risk adjustment," Dr. Lindenauer and his associates said.
As expected, the reductions in inpatient hospitalizations for the three control conditions were significantly smaller than those for pneumonia hospitalizations. Ischemic stroke, STEMI, and ruptured aortic aneurysms were indeed "less susceptible to secular changes in the choice of an alternative principal diagnosis," they pointed out.
Also as expected, the proportion of pneumonia patients discharged to nursing facilities and hospices did not account for the large decline in pneumonia inpatients.
The results of the primary analysis in this study were supported by those of a secondary analysis of bacteriologic types. Hospitalization rates for pneumococcal, pseudomonas, and staphylococcal pneumonias all declined to a similar extent as overall pneumonia and were offset by matching rises in the rates of sepsis due to these organisms.
The study findings have important implications well beyond the scope of pneumonia. "Several recent studies have reported very rapid growth in the rate of hospitalizations of patients with sepsis and severe sepsis, suggesting that the phenomenon in this study" may extend to many other infectious diseases, the investigators said.
Turning to the question of why clinicians might be switching from a principal diagnosis of pneumonia to a principal diagnosis of sepsis/secondary diagnosis of pneumonia, Dr. Lindenauer and his colleagues offered two possible explanations.