 |
|
|
Patient Safety IndicatorsPatient Safety Indicators (PSIs) developed by the Agency for Healthcare Research and Quality[1] are measures that screen for potentially preventable complications during hospitalization. Rates of these complications for a single year at a hospital can provide a snapshot view of performance. Trending these indicators over time affords the ability to assess whether observed differences are likely the result of normal variation (i.e., rates differ from average in a single year) or if observed differences reflect a pattern (i.e., rates differ from average for longer periods of time). This longitudinal perspective may identify models for excellence or areas that could benefit from quality improvement initiatives.
Last year’s Quality Report provided comparisons of Vermont PSI rates to the northeast and national statistics. The report also highlighted specific hospitals that were significantly above or below the state average rate. This year, the Quality Report incorporates another year of data (2007) to provide an update on these measures.
It is important to keep in mind that these measures identify rare events by design. Trending these QIs is particularly important because sporadic events can make performance appear erratic even though there is no statistical difference. This phenomenon is further magnified by smaller denominators at many Vermont hospitals.
General Findings
The addition of 2007 data to the PSIs revealed the following overall trends:
- Vermont rates do not differ significantly from national rates for the PSIs considered;
- Obstetric trauma (with and without instrumentation) rates in Vermont have decreased considerably since 2002. Northwestern Medical Center, Rutland Regional Medical Center, Fletcher Allen Health Care, and Southwestern Vermont Medical Center all experienced a decreased incident rate; and
- There have been no events of transfusion reaction during the study period (2002 – 2007).
Recommendations
These indicators are most suited to “flag” or identify potential quality issues or patterns of excellence. For providers and/or communities with trends of less than favorable performance on these measures they may provide direction for future research and examination of the provider community as well as external influential factors such as patient support system, socio-economic and psycho-social factors.
By design these measures identify rare events. Trending these QIs is particularly important because sporadic events can make performance appear erratic even though there is no statistical difference. This phenomenon is further magnified by smaller denominators at many Vermont hospitals.
These measures should continue to be monitored. Hospitals experiencing trends of increasing rates should conduct further investigation for possible explanations and evaluation of the quality of care.
Measure-specific Findings
Accidental Puncture or Laceration
Birth Trauma – Injury to Neonate
Death among Surgical Inpatients (Failure to Rescue)
Iatrogenic Pneumothorax
Obstetric Trauma – Vaginal with instrument
Obstetric Trauma – Vaginal without instrument
Postoperative Hip Fracture
Postoperative Pulmonary Embolism or Deep Vein Thrombosis
Postoperative Wound Dehiscence
Death in Low-Mortality DRGs
Foreign Body left during Procedure
Transfusion Reaction
Data and Methodology
The AHRQ indicators use administrative hospital discharge data as a window into the delivery of medical care. These data, which are collected as a routine step in the delivery of hospital services, provide information on diagnoses, procedures, age, gender, admission source, and discharge status. Although administrative data alone cannot provide a complete assessment of quality, they can be used to identify potential quality problems or “flag” areas of excellence which can then be further investigated and studied, either to address needed improvements or to assist adoption of excellence at other facilities or in other communities.
The 2009 Quality Report presents analyses of hospital discharge data from 2002 - 2007. This extends and updates the measures reported in last year’s Quality Report. Incorporating the additional year of data (2007) allows further examination of trends.
|
|
|
|
