1.    Pressure Ulcer Prevention and Treatment in Hospitals. A PBE study conducted at a hospital in the Rocky Mountain West to determine if development of pressure ulcers (PrUs) is related to the severity of illness of the patient, and if implementation of a protocol to stabilize the process of care for prevention of PrUs reduces the incidence of PrUs. 

Population:  2,500 at-risk patients/2,500 admissions from 1992 to 1993 in a tertiary-care hospital in the Rocky Mountain West.

Example of Findings: Patients who developed PrUs were sicker on average on admission than patients at risk who did not develop PrUs.  Following implementation of the PrU prevention protocol, PrUs decreased 300 per year, saving on average $1.268 million per year.

2.    Managed Care Outcomes Project.  A PBE study was conducted at six HMOs throughout the US to determine the effect of HMO organizational strategies on outpatient visits, hospital admissions, and prescriptions for patients with at least one of five diseases (arthritis, asthma, epigastric pain/ulcer, hypertension, and otitis media).

Population:  13,158 patients/90,158 office visits/1,871 hospital admissions from 1992 to 1993 (patients followed for one year) in six HMOs throughout the US.

Example of Findings: For the five study diseases, physician continuity was associated with reduced healthcare utilization.  The findings lend credence to the hypothesis that physician continuity, combined with chart and site continuity, is associated with reduced health care costs.

3.    Formulary Limitations and the Elderly: Results from the Managed Care Outcomes Project. A PBE study was conducted at six HMOs throughout the US to determine whether restrictive formulary limitations on drugs are associated with differences in health care resource utilization, including office visits, prescriptions, and hospitalizations, and whether this association varies by age.

Population: 13,000 patients from six HMOs throughout the US were assessed for severity of illness at every encounter for one year, totaling more than 99,000 outpatient visits, 480 emergency department visits, 1,000 hospitalizations, and 240,000 prescriptions.

Example of Findings: Greater limitations in drug classes were significantly associated with greater health care resource utilization; these associations were sometimes significantly greater for elderly patients after controlling for severity of illness and other variables.

4.    Intended and Unintended Consequences of HMO Cost Containment Strategies: Results from the Managed Care Outcomes Project. A PBE study was conducted at six HMOs throughout the US to determine the effect of HMO cost-containment strategies on outpatient visits, hospital admissions, and prescriptions on patients with at least one of five diseases (arthritis, asthma, epigastric pain/ulcer, hypertension, and otitis media).

Population: 13,000 patients from six HMOs throughout the US were assessed for severity of illness at every encounter for one year, totaling more than 99,000 outpatient visits, 480 emergency department visits, 1,000 hospitalizations, and 240,000 prescriptions.

Example of Findings: For four study diseases (but not for otitis media), greater levels of formulary limitations on drugs to treat these diseases were associated significantly with higher rates of emergency department visits and hospitalizations and positively (but not always significantly) with drug cost, drug count, and office visits.

5.    Abdominal Surgery.  A PBE study was conducted at eight hospitals to determine areas of significant improvement in the management of abdominal surgery patients. 

Population:  1,007 patients/1,007 admissions from 1994 to 1997 in eight hospitals in the Pacific West.

Examples of Findings:  After controlling for disease severity, patients with parenteral or enteral feeding that occurred within 48 hours of surgery and met at least 60% of the patient's protein and calorie goals had lower lengths of stay and hospital costs.  The study also found the use of PCA pump to be significantly associated with a higher rate of post-operative surgical wound infection, 10.7% for PCA users vs. 4.0% for non PCA users.

6.    Pediatric Asthma and Bronchiolitis.  A PBE study to improve treatment of asthma and bronchiolitis and to measure pediatric severity of illness generally.

Population:  Severity of illness data for 13,884 patients/16,557 admissions from 1995 to 1996 in 10 pediatric hospitals throughout the US.  PBE data for 762 asthma patients and 782 bronchiolitis patients.

Example of Findings: Extensive variation in practice was found across all 10 pediatric hospitals regardless of severity of illness.  If the best practices were used in all the study institutions we would expect on average a 0.5 day decrease in length of stay and a $1,679 decrease in cost per case for pediatric asthma. For pediatric bronchiolitis, we would expect on average a 2.0 day decrease in length of stay and a $4,982 decrease in cost per case.

7.    Coronary Artery Bypass Graft Surgery. A PBE study to improve surgical processes and CABG outcomes.

Population:  2,241 patients/2,243 admissions from 1994 to 1996 in 15 hospitals throughout the US.

Example of Findings: After controlling for severity of illness, several treatments were associated with longer length of stay, higher cost, larger increase of severity of illness from admission (first 24 hours) to maximum during the hospital stay, and more post-operative infections.

8.    Bone Marrow Transplant and Colorectal Cancer.  A PBE study to improve nutritional support during chemotherapy and radiotherapy for bone marrow transplant or colorectal cancer patients. 

Population:  352 bone marrow transplant patients/443 admissions from 1995 to 1997 in four tertiary-care hospitals in the Northeast and Southwest and 217 colorectal cancer patients/137 admissions/791 visits from 1992 to 1996 in four hospitals in the Northeast, Midwest, and Pacific Southwest.

Example of Findings: Special nutritional supplements were associated with fewer complications and better survival.

9.    Congestive Heart Failure. A PBE study was conducted to determine areas of significant improvement in the management of congestive heart failure (CHF) patients.

Populations:  215 patients/338 admissions/791 visits from 1995 to 1996 in a hospital in the Pacific Northwest and 116 patients/384 admissions/316 visits from 1994 to 1996 in a hospital in the mid-Atlantic region.

Examples of Findings: Longer length of stay was associated with drug underdosing. A greater number of admissions was associated with not giving education, as well as with noncompliance with diet by patients.

10. National Pressure Ulcer Long-Term Care Study.  A PBE study to improve prevention and treatment of pressure ulcers in nursing homes.   Further analyses include best care for nutrition deficiencies, incontinence, dementia, and nurse staffing.

Population:  2,486 residents/6,833 severity assessments from 1996 to 1997 in 109 facilities belonging to six long-term-care provider systems throughout the US.

Examples of Findings: Disposable briefs, supplement use, and combination medications are significantly associated with development of fewer pressure ulcers, which results in less resident suffering and decreases costs in nursing homes.  Registered nurse direct care time >0.5 hours/resident/day was associated with fewer PrUs, urinary tract infections, and hospitalizations with a cost savings to society of $3,191 per resident per year.

11. Falls in Hospitals.  A PBE study to improve prevention of falls in hospitals.

Population:  280 patients/285 admissions from 1995 to 1996 in a hospital in the Pacific Northwest.

      Examples of Findings: Fall patients stayed an average of 2.8 days longer than non-fall patients. Older patients, those with restraints, those using nicotine substitute, and those with medical equipment were less likely to fall.  Patients who were sicker, used glasses or contacts, used non-narcotic analgesics, had a secondary psychiatric diagnosis, climbed out of bed, or used a walker or wheel chair during hospitalization were more likely to fall. 

12. Abdominal Surgery Complications.  A PBE study was conducted to determine areas of significant improvement in the management of abdominal surgery patients

Population: 977 patients/977 admissions from 1997 to 1998 in three hospitals in the Rocky Mountain West.

Examples of Findings: After controlling for disease severity, several treatment variables under provider control including timing and amount of nutritional support, use of PCA pump to control pain, and use of drains were associated with greater increase in severity from admission (first 24 hours) to maximum during the hospital stay, longer length of stay, higher cost, and higher infection rates.

13. Optimal Ambulatory Type 2 Diabetes Care.  The purpose of this PBE study was to determine interventions associated with significant improvement in the management of adult type 2 diabetes.  The study was conducted at four West Coast ambulatory clinics between March 1999 and December 2002. 

Population:  Detailed patient, treatment, and outcome data were collected on 228 adult type 2 diabetes patients 35-65 years old.  The patients had 3,069 diabetes care visits to 65 physicians. Patients were followed for 3 years. 

Example of Findings: After controlling for severity of illness, self-monitoring of blood glucose along with consistent provider discussion was strongly associated with better serum glucose control and with fewer hospitalizations. The findings from this study of adult outpatients with type 2 diabetes in actual practice of care confirm findings from randomized trials of similar patients.  These indicate that intensive therapy is an effective way to reach the major goals of diabetes treatment:  lowering blood glucose to appropriate levels and avoiding and postponing the onset of serious complications.

 

14. Telemedicine Depression Study.  Study to discover impact of telemedicine to treat depression in patients in rural areas.

 

Population:  998 patients/4,273 clinic visits from 1999 to 2000 in rural settings in the Midwest.

 

15. Telemedicine Dermatology Study.  Study to discover impact of telemedicine to treat dermatology patients in rural areas.

 

Population:  343 patients/779 visits from 1999 to 2000 in rural settings in the Midwest

 

16. IV Heparin Study.  The purpose of this study was to determine patient, environment, and treatment factors that are associated with significantly lower rates of intravenous (IV) heparin medication errors, and to discover optimal ways to prevent IV Heparin medication errors.

 

Population:  426 patients from 2001 to 2003 in six hospitals in the Pacific West.

 

Examples of Findings: 63.9% of heparin orders were written in Physician Orders, 13.2% were in facility protocols, 13.1% were in specialized disease protocols such as for Acute MI, and the remaining 9.8% were written in ER orders.  Only 82.1% of patients are weighed on the first day of hospital admission. Clinicians have not agreed on a standard method to determine a heparin bolus dose or infusion rate across facilities.  Patients with reported medication errors and/or abnormal PTT values had longer lengths of stay by approximately two days although there was not a significant difference in severity of illness.  Patients who are under anticoagulated (PTT<50) had longer lengths of stay than patients who were over anticoagulated (PTT>120).  Heparin protocols are not used frequently in practice although they have been shown to improve anticoagulation.  The resistance by physicians to use protocols perpetuates an environment that allows medication errors to be prevalent in acute care settings.  The development of efficient and effective heparin protocols and adherence by clinical staff has the potential to improve patient care and safety and greatly reduce care costs by decreasing length of stay.

 

17. National Hospice Outcomes Project.  A PBE study to discover optimal pain control, dyspnea control, and self-determined life closure methods for hospice patients.

Population:  1,306 patients from 2001 to 2003 in 13 hospices throughout the US.

Example of Findings: The NHOP analyses demonstrate that final opioid dose, but not percent change in dose, was one of several factors associated with survival, but the association was weak; in multivariate analyses, this and other relevant factors explain only a very small percentage of variation in survival. The implication that opioid dose poses an extremely small risk of hastened death in this population was supported further by the relatively long intervals between final dose change and death, and the lack of higher opioid risk in subsamples receiving high doses. This analysis suggests that the timing of death in populations with far advanced illness involves a complex interplay of variables, including important factors that were not assessed in this study, and that opioid therapy should not be the focus of future research of this type.

18. C-Section Study.  A PBE study to discover optimal labor and delivery methods to reduce C-section rates and prevent labor and delivery complications.

Population:  1,343 labor and delivery patients from December 2001 and January 2002 in 8 southwest Texas hospitals.

Examples of Findings:  In comparing rates of postpartum complications with vaginal vs. C-section deliveries, there was a trend of fewer complications when the C-section was scheduled.  When looking at gender of delivering physician, the proportion all C-section vs. vaginal deliveries was significantly higher for female delivery physicians compared to males.  The proportion of unscheduled C-sections was significantly higher for female delivering physicians.  There was no significant difference in female delivering physicians for scheduled C-sections.

 

Female delivering physicians had significantly fewer years of experience on average (8.1 years vs. 21.9 years, p<0.0001) and significantly fewer law suits on average (1.0 vs. 3.4, p<0.0001).  To control for years of experience, we did a sub-analysis of physicians with <15 years experience.  We found the unscheduled C-section rate for female delivering physicians was 42% compared to 13% (p=0.0009) for male delivering physicians. The scheduled C-section rate for female delivering physicians was 24% compared to 18% (p=0.62) for male delivering physicians. The average number of law suits for female delivering physicians was 1.0 compared to 2.0 (p<0.0001) for male delivering physicians. 

 

In determining if there was an association with the time of starting prenatal care and C-Section deliveries, the study found that, for scheduled and unscheduled C-sections combined, patients delivering by C-section had borderline significantly younger average gestational age at start of prenatal care than those who delivered vaginally (12.6 weeks vs. 14.6 weeks, respectively, p=0.06).  Patients delivering by C-section were also significantly older at delivery than those who delivered vaginally (30 years vs. 26.7 years, p=<.0001).

19. Post-Stroke Rehabilitation Outcomes Project.  A PBE study to discover optimal interventions in post-stroke rehabilitation.

Population:  1,291 post-stroke rehabilitation patients from 2001 to 2003 in 7 inpatient rehabilitation facilities (6 across the US and 1 in New Zealand).

Examples of Findings:  Fewer days from stroke symptom onset to rehabilitation admission (earlier rehabilitation admission) is significantly associated with higher total, motor, mobility, and ADL discharge FIM scores, controlling for severity of illness and initial functioning.  Tube feeding severe stroke patient more than 25% of their stay is significantly associated with higher total, motor, and cognitive FIM, and greater improvement of severity of illness by discharge, even though tube fed patients had a higher pneumonia development rate, which is a known complication of tube feeding.  More time spent in high-level rehabilitation activities, such as gait, upper extremity control, and problem solving in the first three hours of therapy for low-level stroke rehabilitation patients is significantly associated with higher total, motor, and cognitive FIM at discharge.

20. Joint Replacement Outcomes in IRFs and Nursing Treatment Sites.  A PBE study to discover optimal interventions in hip and knee joint replacement rehabilitation.

Population:  2,158 hip and knee replacement rehabilitation patients from February 2006 through February 2007 in 8 freestanding skilled nursing facilities, 1 hospital-based SNF, and 11 inpatient rehabilitation facilities.

Examples of Findings: Hip and knee replacement patients in hospital-based rehabilitation centers, commonly known as inpatient rehabilitation facilities (IRFs), recover faster and have marginally better functional outcomes compared to patients who receive their rehabilitation in skilled nursing facilities (SNFs) where patients typically stay longer but receive less intense services. Apart from setting, the study found three things that are associated with better outcomes:

·         Starting rehabilitation earlier rather than later after surgery,

·         Obtaining more intense therapy, namely more physical and occupational therapy per day akin to what is found in IRFs,

·         Going to a facility that serves a ‘medium volume’ of joint replacement patients. The study’s medium-volume facilities, both SNFs and IRFs that treated between 100 and 183 joint replacement rehabilitation patients yearly, had the best results.  Low-volume facilities did much less well.

21. Analysis of Staff Workflow and Associated Costs Among Green House and Traditional Nursing Home Sites. Study to analyze front-line caregiver daily workflow in skilled nursing settings, including Green House (GH) homes, and assess if Green House staffing was more expensive than traditional nursing home staffing. 

Population: Hundreds of residents and staff in 27 sites:  13 nursing units in Traditional skilled nursing facilities and 14 Green House homes across the US.

Examples of Findings: Staffing from nursing and non-nursing departments combined, excluding administrative staff, was 0.3 hours per resident per day (18 minutes) less in GH homes compared to Traditional SNFs. CNAs in GH homes, while responsible for more non-nursing activities like laundry and housekeeping, spent 0.4 more hours per resident per day (24 minutes) in direct care activities compared to CNAs in Traditional SNFs. The results challenge the assumption that staffing efficiencies cannot be achieved in small environments like a GH home. While GH model has higher ratio of CNA staff to residents compared to Traditional SNF units, overall staff time (combined total of nursing and non-nursing HPRD) is slightly less in GH homes. Compared to Traditional SNFs, the GH model allows for expanded responsibilities of CNAs in indirect care activities and more time in direct care activities and engaging directly with resident.

22. Analysis of Changes in ADL Assistance Levels in Traditional Nursing Homes and THE GREEN HOUSE® Project Sites. The principal objectives of this study were to conduct an analysis of ADL change among Green House elders compared to residents in the traditional nursing home, and gain an understanding of Green House model ADL outcomes compared to those in traditional nursing homes.

Population: A total of 4 Green House organizations participated in the study.  From these organizations, 13 sites (GH homes and Traditional SNF units) were included in analyses.  The total sample size was 255 residents: 97 GH residents and 158 Traditional SNF residents.

Example of Findings: After controlling for resident characteristics known to be associated with ADL decline, there was significantly less decline in ADL assistance levels at 9- and 15-month follow-up time periods for GH residents compared to residents in Traditional SNFs. 

23. Traumatic Brain Injury Outcomes Project.  A PBE study to discover optimal interventions in post-traumatic brain injury rehabilitation. Study in progress.

Population:  2,315 traumatic brain injury rehabilitation patients from October 2008 through August 2011 in 10 inpatient rehabilitation facilities (9 across the US and 1 in Canada).

24. Spinal Cord Injury Outcomes Project.  A PBE study to discover optimal interventions in spinal cord injury rehabilitation. Study in progress.

Population:  1,500 spinal cord injury rehabilitation patients from September 2007 through December 2009 in 6 inpatient rehabilitation facilities across the US.

25. Pressure Ulcer Prevention in Spinal Cord Injury.  A PBE study to discover optimal interventions to prevent pressure ulcers in spinal cord injury patients in hospitals and rehabilitation. Study in progress.

Population:  900 spinal cord injury patients from September 2007 through March 2010 in 1 hospital and 1 inpatient rehabilitation facility in the eastern US.

26. Turning Frequency in Nursing Home Residents.  A PBE study to determine the impact of turning at 2, 3, or 4 hour intervals for frail elderly nursing home residents. Study in progress.

Population:  900 frail nursing home residents from March 2008 through December 2010 in nursing homes across the US and Canada.

27. Pressure Ulcer Healing Outcomes Project.  A PBE study to determine optimal treatments for pressure ulcers at various stages. One of the outputs of this project in an On-Time quality improvement manual that describes how to prevent pressure ulcers.  Study in progress.

Population:  Approximately 2,600 frail nursing home residents with approximately 5,000 pressure ulcers from December 2007 through September 2010 in nursing homes across the US.

28. Preventing Injurious Falls in Nursing Homes.  Study to implement fall prevention strategies in nursing homes. Study in progress.

29. Genetic Risk Factors for Severe Influenza in Children.  Study to determine genetic factors that are associated with severe influenza in children. Study in progress.

30. Comparative Effectiveness of Interventions for Chronic Pain Management.  Study to determine the combinations of medications, surgery types, and ancillary interventions that are associated with best pain management for patients with chronic cancer or non-cancer pain. Study in progress.

Population: Several thousand patients with chronic cancer or non-cancer pain seen in four pain clinics located at Cornell Weill Medical Center, Memorial Sloan Kettering Cancer Center, and Hospital for Special Surgery in New York City.

31. Severity of Illness Measurement Using the Comprehensive Severity Index.

Population:  14,324 patients/15,853 admissions from 1991 to 1992 in 13 ICUs in the Rocky Mountain West.

Population:  10,856 patients/13,962 admissions from 1993 to 1996 in a tertiary-care hospital in the Pacific West.

Population:  8,014 patients/9,554 admissions/51 visits from 1991 to 1995 in a tertiary-care hospital in the Rocky Mountain West.

Population:  3,490 patients/3,487 admissions/492 visits from 1993 to 1997 in a hospital system in the Midwest.

Population:  480 patients/480 admissions from 1993 to 1995 in a psychiatric hospital in Eastern Canada.