Development and validation of hospital "end-of-life" treatment intensity measures

Abstract

Background: Health care utilization among decedents is increasingly used as a measure of health care efficiency, but decedent-based measures may be biased estimates of care received by "dying" patients.

Objective: To develop and validate new measures of hospital "end-of-life" treatment intensity.

Research design: Retrospective cohort study using Pennsylvania Health Care Cost Containment Council (PHC4) discharge data (April 2001-March 2005) and Centers for Medicare and Medicaid Services (CMS) data (January 1999-December 2003).

Subjects: Patients 65 and older admitted to 174 Pennsylvania acute care hospitals.

Measures: Hospital-specific standardized ratios of intensive care unit (ICU) and life-sustaining treatment (LST) use among terminal admissions (decedents) and admissions with a high probability of dying, and spending and use of hospitals, ICUs, and physicians among patients in their last 6 months of life.

Results: There was marked between-hospital variation in the use of the ICU and LSTs among decedents and admissions with high probability of dying. All hospital decedent and high probability of dying measures were highly correlated (P < 00001). In principal components factor analysis, all 4 of the last-6-months cohort-based measures, the decedent and high-risk admission-based ICU measures, and 8 of the 12 decedent and high probability of dying LST measures loaded onto a single factor, explaining 42% of the variation in the data.

Conclusions: Hospitals' end-of-life intensity varies in the use of specific life-sustaining treatments that are somewhat emblematic of aggressive end-of-life care. End-of-life intensity is a relatively stable hospital attribute that is robust to multiple measurement approaches.

Figure 1. Frequency of Predicted Probability of Death among all Admissions (Panel A) and among Terminal Admissions (Panel B) Age 65 and Older, Pennsylvania 2001–2005

The histogram on the left (Panel A) represents all admissions among Pennsylvanians over age 65 over the 4 study years that fall within particular bands of predicted probability of death upon admission, based upon the MediQual Atlas proprietary risk model. Admissions in the top 5% of risk of death of admission, corresponding to a predicted risk of 21% or higher, make up the high probability of dying admission group. The histogram on the right (Panel B) represents the total number of terminal admissions among Pennsylvanians over age 65 that fall within particular bands of predicted probability of death upon admission. Many terminal admissions occur among persons who had relatively low predicted probability of death upon admission. Thirty-nine percent of terminal admissions were among patients in the high probability of dying group.

Figure 2. Crude ICU Admission Rate by Deciles of Admission Risk of Death, 4 Pennsylvania Hospitals and the State Average, 2001–2005

This graph depicts the crude ICU admission rate among patients over age 65 along the y-axis, against the patient’s predicted probability of death upon admission, categorized in deciles of risk along the x-axis, for two Pittsburgh hospitals and two Philadelphia hospitals, compared to the state average (dotted line).

Figure 3. Correlation between Retrospective and “Prospective” PHC4 Indices of Hospital End-of-life Intensity, Pennsylvania 2001–2004

The figure is a plot of 172 hospitals. Each bubble represents one hospital and the size of the bubble is proportional to the number of admissions among patients over age 65 during the 4 years of observation. The hospital’s index of intensity among high probability of dying (prospective “end of life”) admissions is represented along the y-axis and its factor-based index of intensity among decedent (retrospective “end of life”) admissions is represented along the x-axis. The index is a factor score of 6 underlying standardized (observed-to-expected) ratios: ICU admission, ICU length of stay, intubation/mechanical ventilation, tracheostomy, hemodialysis, feeding tube placement. Higher values reflect higher intensity. The cross-hatches represent 95% confidence intervals for each index value.

Figure 3. Correlation between Retrospective and “Prospective” PHC4 Indices of Hospital End-of-life Intensity, Pennsylvania 2001–2004

The figure is a plot of 172 hospitals. Each bubble represents one hospital and the size of the bubble is proportional to the number of admissions among patients over age 65 during the 4 years of observation. The hospital’s index of intensity among high probability of dying (prospective “end of life”) admissions is represented along the y-axis and its factor-based index of intensity among decedent (retrospective “end of life”) admissions is represented along the x-axis. The index is a factor score of 6 underlying standardized (observed-to-expected) ratios: ICU admission, ICU length of stay, intubation/mechanical ventilation, tracheostomy, hemodialysis, feeding tube placement. Higher values reflect higher intensity. The cross-hatches represent 95% confidence intervals for each index value.

Appendix Figure. Correlation between Retrospective and “Prospective” PHC4 Measures of Hospital End-of-life Intensity, Pennsylvania 2001–2004

Each of the 8 panels is a plot of individual hospitals with each dot representing one hospital. The hospital’s standardized (observed-to-expected) ratio among high probability of death (prospective “end of life”) admissions is represented along the y-axis and its standardized ratio among decedent (retrospective “end of life”) admissions is represented along the x-axis.

Appendix Figure. Correlation between Retrospective and “Prospective” PHC4 Measures of Hospital End-of-life Intensity, Pennsylvania 2001–2004

Each of the 8 panels is a plot of individual hospitals with each dot representing one hospital. The hospital’s standardized (observed-to-expected) ratio among high probability of death (prospective “end of life”) admissions is represented along the y-axis and its standardized ratio among decedent (retrospective “end of life”) admissions is represented along the x-axis.