Monitor - January 1998

A quarterly publication for members of the Medical Outcomes Trust
January 1998

Evaluating Outcomes of Toal Hip &
Total Knee Replacements
Editor's Note:
Thanks to the following individuals whose contributions made this article possible: Robert B. Keller from Maine Medical Assessment Foundation; Deborah Kuhr, Dominic Cioffi and Alan Praemer from the American Academy of Orthopedic Surgeons; Stephen Herrington from Joint Implant Surgeons; Norman Johanson from Temple University; and Fred Andersen from University of Massachusetts Medical Center: Center for Outcomes Research.

Initiatives
Summaries of efforts to evaluate outcomes of hip and knee replacements

This is the sixth issue of the Monitor to explore the evaluation of patient outcomes related to a particular clinical topic. While the previous issues have featured clinical conditions, e.g., asthma, congestive heart failure and diabetes, in this edition we examine the state-of-the-field of patient outcomes related to an intervention, IE, lower extremity total joint replacements (TJR), specifically, total hip replacements (THR) and total knee replacements (TKR).

In 1994, 124,000 total hip replacement surgeries were performed in the U.S., an increase from 119,000 in 1990. As the number of total hip procedures performed in 1994 increased four percent from 1990, the average length of hospital stay per procedure decreased 33.3 percent from 10.8 to 7.2 days. During the same four years, the number of total knee replacements increased a significant 62 percent from 129,000 in 1990 to 209,000 in 1994, and the average length of stay decreased 35 percent from 10.5 to 6.8 days. While this information is important and necessary to understanding the value of our health care dollars, it is not sufficient. Information about the change in patients' health status from before to after surgery is also necessary to help us determine the amount of health we are buying with our health care dollars.¹

In developing this issue of the Monitor we have learned that the field of patient-based health outcomes of total joint replacements is more advanced than the work being done in many of the other clinical topics we have explored to date. The experts we spoke with discussed five reasons that the development and use of health outcomes of joint replacements has progressed to its current status.

To begin with, practical concerns related to the changing face of health care²put the study of health outcomes of TJR on the agenda. When compared with many other surgical procedures, total hip and knee replacements are considered elective procedures, IE, they are not life saving. Like all elective surgery for musculoskeletal conditions, there is a great deal of variation in the rate with which these procedures are performed. In the 1980's, the number of total hip and knee replacement surgeries that were being performed was increasing significantly and they were associated with rapidly escalating health care costs. Extensive variation in the performance of these procedures led to the question - what is the right rate? Leaders in the field sought to generate evidence of the value (defined as quality divided by cost) of TJR. Patient-based outcomes were known to be one way to quantify health care quality.

There are other reasons that the field of health outcomes of TJR has progressed to the extent that it has today. First, the basic problems for which joint replacements are performed are joint pain and functional disability, e.g., difficulty walking and performing routine activities of daily living. This is a key issue meaning that it is natural and routine for patients and physicians dealing with TJR to focus on physical function, pain and other measures of patient-based health status. Second, the time window within which health status and functional changes can be observed and measured following total hip and knee replacements is relatively brief, e.g., six to 12 months.

Third, changes in health-related quality of life and patient satisfaction following TJR tend to be significant. For example, one study of 54 adults following total hip arthroplasty used the SF-36³ and demonstrated: 1) all eight SF-36 scales improved a statistically significant amount following surgery, the Physical Component Summary (PCS) scale improving twice as much as the Mental Component Summary (MCS) scale; 2) the average improvement of the PCS score was 9.5 (from 39 to 48.5) representing an improvement in physical health large enough to move from the 15th to the 33rd percentile of the distribution of scores in the general U.S. population (normative data); 3) 87 percent fewer people had to reduce their work three months after hip replacement as compared with before surgery; and 4) the number of individuals who had severe pain declined 90 percent after surgery.⁴

For all of these reasons, it is easier for many physicians to connect TJR with the idea of patient-based health status assessment because both the underlying clinical problems and the intervention lend themselves to it more naturally than other clinical conditions that are not perceived to be about function and well-being.

However, there is another important reason this field has made the significant strides it has - strong leadership from key professional organizations and individual opinion leaders. One such leader, Robert Keller, MD, of the Maine Medical Assessment Foundation, has argued for nearly a decade that in response to the dramatic changes in health care delivery systems, physicians must move quickly into the field of outcomes research because outcomes data are the "ammunition" that physicians need to demonstrate: 1) their awareness of problems in medicine, e.g., high costs and practice variation, and 2) that they have protocols to address the problems.²

We believe there is much to be learned from the work done to evaluate the outcomes of total hip and knee replacements to date. The MODEMSª Program signifies a major step forward in the field of patient-based health outcomes, not only as it relates to musculoskeletal conditions, but also as a model for other fields of medicine and health care.

The MODEMS™ Program (Musculoskeletal Outcomes Data Evaluation and Management System) was designed to improve the quality and value of musculoskeletal care by providing physicians with tools and methods to collect, organize and analyze information about the outcomes of the health care patients receive. The program is jointly sponsored by the American Academy of Orthopaedic Surgeons (AAOS), the Council of Musculoskeletal Specialty Societies (COMSS), and the Council of Spine Societies (COSS). Together, a Task Force comprised of representatives of the three organizations developed MODEMS™ and continues to oversee it.

Figure 1.
MODEMS™: PARTICIPATING SPECIALTY SOCIETIES Lower Limb American Academy of Orthopaedic Surgeons American Association of Hip and Knee Surgeons American Orthopaedic Society for Sports Medicine Hip Society Knee Society Orthopaedic Rehabilitation Association Orthopaedic Trauma Association Arthroscopy Association of North America Mexican Orthopaedic Foot and Ankle Society Musculoskeletal Tumor Society Pediatrics American Academy of Orthopaedic Surgeons Pediatric Orthopaedic Society of North America American Academy of Pediatrics Shriner's Hospitals Spine American Academy of Orthopaedic Surgeons Council of Spine Societies North American Spine Society Scoliosis Research Society Cervical Spine Research Society Orthopaedic Rehabilitation Association American Spinal Injury Association Arm, Shoulder and Hand American Academy of Orthopaedic Surgeons Institute for Work and Health, Toronto American Association for Hand Surgery American Orthopaedic Society for Sports Medicine American Shoulder and Elbow Surgeons Orthopaedic Rehabilitation Association Arthroscopy Association of North America American Society of Plastic and Reconstructive Surgeons

Since the early 1990's, the AAOS with support from COMSS and COSS has worked with many specialty societies to develop and implement health-related quality of life data collection instruments for four subgroups of musculoskeletal care - pediatrics, the spine, the upper extremity (the Disabilities of the Arm, Shoulder and Hand Instrument or DASH) and the lower limb. Figure 1 demonstrates the collaboration in which approximately 23 specialty societies engaged to create this system. Four workgroups created a series of "core" outcomes data collection instruments with the aim of defining a generic core of information to be collected from all patients with a specific condition in each of the four areas.⁵

The generic core of information that is collected with the spine, upper extremity and lower limb MODEMS™ instruments is the same. It includes: 1) primary and secondary diagnoses (ICD-9 codes), and procedures (CPT codes), 2) demographic and other identifying information, 3) a description/illustration of locations and intensity of pain, 4) comorbities, and 5) the SF-36. This generic core of information is supplemented by modules of questions that gather information about the three broad categories of musculoskeletal conditions, IE, spine, upper extremity and lower limb.⁵

The MODEMS™ Lower Limb Outcomes Instrument includes: 1) core items that are collected for all patients with lower limb problems and 2) three modules which are used selectively based on the joint in which the patient has a problem and the type of problem. The three specific modules are:

Thus, orthopedists who use MODEMS™ collect the generic core of information on all of their patients with lower extremity problems , the core lower limb items and the items in one of the three specific modules listed above. For individuals undergoing total hip or knee replacements, the hip and knee module would be used. This design of the lower limb instrument ensures that standardized, comparable information is collected for all patients with lower limb problems plus items specific to their particular problem.

The American Association of Hip and Knee Surgeons (AAHKS), in collaboration with the AAOS, developed a methodology to collect process information for patients undergoing hip and knee replacements, MODEMS-HK. The information collected with MODEMS-HK is captured in five modules specifically: 1) Post Discharge Complications Form, 2) Discharge Form, 3) Hip-Operation Form, 4) Knee-Operation Form, and 5) Physical Exam Form.⁶

The unique features of MODEMS-HK were developed by the Outcomes Task Force of the AAHKS to facilitate the development of an electronic medical record so that outcomes assessment could be incorporated in the office routine. Letters to referring physicians, operative notes and discharge summaries can all be generated from the data using software developed with the appropriate functionality.

The collaborators who developed the MODEMS™ Lower Limb Instrument selected an initial set of items on the basis of their clinical experience and information in the published literature. The items were selected to reflect function specific to the lower limb that was not covered by the SF-36. Two types of research and evaluations were conducted as part of the instrument development process including: 1) research into the reliability and validity of the tool, and 2) a pilot study which evaluated its use in a practical context. (The research and evaluation described here refers to the MODEMS™ Lower Limb Outcomes Instrument and does not include the process information collected with the MODEMS-HK.)

Research was conducted to establish the reliability and validity of the instrument including its sensitivity to change. A final report of this research was recently submitted to the AAOS which will be making the findings public in the near future. However, a summary of the findings includes: 1) the scales in the Lower Limb Instrument measure a wide range of pain and function,

2) they are generally highly reliable, 3) they are correlated with other measures of physical function (evidence of validity), and 4) they are sensitive to change. The study's conclusion was that, combined with the SF-36, the scales in the Lower Limb Outcome Instrument should prove a useful tool in the measurement of outcomes of lower limb orthopaedic patients.⁷

In addition to the research of the psychometric properties of the tool, a three month pilot study of MODEMS™ was conducted to evaluate its practical utility. The conclusions of the study included: 1) collecting outcomes data in a systematic and consistent way and transmitting the data to a central database was feasible, 2) over 90 percent of the practices that provided data deemed the experience "worth the time and costs", and 3) patients did not object to providing the information.⁵

Based on the lessons learned from the pilot study, in February, 1996, the AAOS Board of Directors approved Phase II of the project, a two-year, $1.1 million initiative.⁵ This program has been underway since the spring of 1997. To date, 32 clinics and 130 orthopaedic surgeons are participating and approximately 3,000 patient records have been transmitted to the national database.

Joint Implant Surgeons (JIS) in Columbus, Ohio is an orthopaedic practice with extensive experience collecting patient-based health status information. Principals at JIS expect the benefits of participating in the MODEMSª Program to include the following. 1) Principals at JIS believe that taking part in MODEMSª at this stage allows them to have input in creating the future standards for outcomes in hip and knee arthroplasty, 2) Participating in MODEMS™ will enable JIS to compare their outcomes with other orthopedists who perform hip and knee replacements and will provide them with information that will inform a process of continuous improvement by iteratively pointing out their strengths and weaknesses, and 3) JIS principals believe that patient-based health-related quality of life outcomes provides them with information to demonstrate the value of TJR to third party payers and purchasers. JIS is investing in patient-based outcomes, recognizing that the payoff may take some time to manifest. They believe that over time the benefits will far outweigh the costs.⁸

References
1. American Academy of Orthopaedic Surgeons. Facts about Total Hip and Total Knee Replacement. May, 1997.
2. Robert B. Keller. "Unstable Knees in Unstable Times." American Journal of Sports Medicine 1996; 24:570-574.
3. JN Katz, MG Larson, CB Phillips, AH Fossel, and MH Liang. "Comparative Measurement Sensitivity of Short and Longer Health Status Instruments." Medical Care 1992; 30:917-925,
4. John E. Ware, Mark Kosinski, and Susan D. Keller. SF-36 Physical and Mental Summary Scales: A User's Manual. Boston, MA. The Health Institute, 1994.
5. "Academy Plans Data Collection System." AAOS Bulletin. April, 1995.
6. "How to Use the MODEMSª Hip/Knee Module and the MODEMS-HK Module (Version 2.0)." American Association of Hip and Knee Surgeons. Rosemont, Illinois. August, 1997.
7. Lawren Daltroy. "AAOS Lower Limb Outcomes Scales: Reliability, Validity and Sensitivity to Change." Final Report to American Academy of Orthopaedic Surgeons. November, 1997.
8. Steven Herrington. Personal Communication. Joint Implant Surgeons. Columbus, Ohio. December, 1997.

The Hip & Knee Registry: A Resource for Measuring and Improving Quality of Care

The Hip & Knee Registry is a surgeon-driven project, established to evaluate outcomes of primary and revision total hip and knee arthroplasty procedures. Initiated three years ago, the registry has accumulated data on 9,121 patients. To date, complete data sets are available on 3,852 hip cases and 3,729 knee cases. Of the more than 900 orthopaedic surgeons throughout the US and Canada enrolled in the registry, 336 have submitted data.

Participating surgeons complete a peri-operative form, and follow-up forms at the first visit after suture removal and annual thereafter. Patients complete a pre-operative form and follow-up forms at the same intervals as surgeons.

Surgeon-completed forms collect information on such variables as diagnosis, coexisting disabling joint diseases, surgical approach, type of implant, antibiotic and antithrombotic prophylaxis, length of stay, discharge disposition, and medical and orthopaedic complications following surgery.

Patient-completed forms collect information on demographics, pre- and post-operative assessment of medical conditions, pre-operative expectations and post-operative satisfaction. Patient forms incorporate a generic measurement of health-related quality of life (the SF-36 Health Survey) and a disease-specific functional assessment instrument (the WOMAC Osteoarthritis Index).

Data is submitted to a centralized data coordinating center, located at the Center for Outcomes Research at the University of Massachusetts Medical Center. The center analyzes data and prepares and distributes individualized quarterly reports to participating surgeons. Data is reported on a North American, regional, and individual basis. Confidentiality of surgeons and patients is maintained through the allocation of confidential identification numbers.

Contact: Fred Anderson, PhD, Director, Center for Outcomes Research at The Hip & Knee Registry helpline: (508) 856 HIPS/4477 or call (888) 459 THKR/8457 for information on how to enroll in the registry.

An Assessment of Clinical and Service Outcomes Pre to Post Operation
for a Total Hip Replacement (THR).

This project is designed to evaluate a range of clinical and service outcomes for patients attending a regional orthopaedic unit for a primary THR. One hundred consecutive patients undergoing surgery for a primary THR were asked to participate in a study assessing changes before and after their operation. Patients were administered the SF-36 and a co-morbidity questionnaire pre-operatively (2 - 5 days). At follow-up (9 - 12 months), patients were again administered the SF-36 and a questionnaire assessing their satisfaction and the quality of care received. Patient records were examined at follow-up to investigate infection, re-admission or revision rates (and reasons for) and complications experienced by patients in the intervening period since surgery.

This investigation will assess changes pre to post operation for patients in areas such as physical functioning and pain, social functioning, role limitations and general health perceptions. These changes will be assessed while controlling for demographics and previous medical history.

Two standardized psychological instruments assessing general self-efficacy and dispositional optimism were included in the pre-operative interview. This information will document where changes occur as a result of a THR and the magnitude of these changes.

Further contact: Dr. Fenton Howell, Specialist in Public Health Medicine, Department of Public Health, North Eastern Health Board, Railway Street, Navan, Co. Meath, Ireland.

The objective of this prospective longitudinal study is to evaluate the extent to which personality effects patients' post operative mood and the putative effect that this may have on surgical outcome. Patients electing to have knee replacement surgery are invited to participate. Pre-operative mood is measured using the Profile of Mood State BiPolar Questionnaire (POMS) and personality is measured using the Eysenck Personality Inventory. Functional ability is assessed pre-operatively and at 1 year post operatively using the American Association of Orthopaedic Surgeons Knee Score and an in house measure called the Eastbourne Hip and Knee Disability Score. Patients' mood is measured post operatively at 3 days, 6 weeks, 4 months and 1 year using the POMS. Demographic data are recorded at initial interview and recruitment into the study. Data analysis will include one way analysis of variance, correlational analysis and categorical analysis where appropriate. The study is currently 3 months underway.

Contact: Dr. Anne Mandy, Senior Lecturer, University of Brighton, 49 Darley Road, Eastbourne BN20 7YR UNITED KINGDOM; phone: +44 127 346 3650; email: Am86@bton.ac.uk

Early work by Mandy & Gard (1995) considered the psychological consequences of surgery from within the immediate post operative period to four months post surgery. Results indicated evidence of Dysphoria at the six week time point. An explanation for such a finding was a perceived difference in outcome. Surgically, at this time point, patients are considered to be fully recovered. However, in terms of functional ability, they may perceive themselves as not having reached full functional recovery.

The objective of this current study is to repeat the earlier study, but on a larger scale and to also include a measure of functional outcome. The current study will measure prospectively, longitudinally the pattern of mood from the pre-operative period to 1 year post operatively. Mood is being measured using the Profile of Mood State BiPolar questionnaire pre-operatively 3 days, 6 weeks, 4 months, 6 months and 1 year post operatively. In addition qualitative data concerning function, pain level and satisfaction with surgical outcome is being recorded from 6 months. Data collection commenced in April and is ongoing. Demographic variables including previous history of depressive disorder is being collected. Analysis will include one-way ANOVA and Chi2.

Contact: Dr. Anne Mandy, Senior Lecturer, University of Brighton, 49 Darley Road, Eastbourne BN20 7UR UNITED KINGDOM, phone: +44 127 346 3650; email: Am86@bton.ac.uk

Total hip replacement (THR) relieves pain for ten to twenty years in over 90% of patients with disabling hip arthritis. These excellent results have been reported almost exclusively from large referral centers, yet over half of primary THR and three quarters of revision THR occur in hospitals which perform fewer than one of these procedures per week in the US Medicare population. We do not know whether the achievements of referral centers have been translated effectively to smaller hospitals where most THR are now performed. The aims of this study are to: (1) document the outcomes of primary and revision THR in a population-based cohort of elderly Americans, including mortality, complications, symptoms, functional status, satisfaction and costs, and (2) identify predictors of outcome, and, in particular, determine whether surgeons and hospitals with a higher volume of THR have better outcomes.

This study employs an innovative design which has never been applied in health services research of a musculoskeletal disorder. We will use Medicare claims data to identify the 80,000 beneficiaries over 65 years old in the US who had primary THR and 15,000 who had revision THR in 1995. Claims data will provide rates of mortality and complications. A random sample of 1,000 primary THR and 1,000 revisions from five states will be selected for detailed study. This includes medical record review to obtain information on the operative procedure, preoperative comorbidity and complications; and patient interviews three years after surgery to obtain information on symptoms, function and satisfaction. Patient-based outcomes will be measured using generic and disease-specific instruments including the SF-12, the WOMAC, and the Harris Hip Score. Data will be obtained by phone or mail, based on the subject's preference. Data on the hospitals and surgeons involved in these cases will be obtained from surveys. Our analyses will focus upon potential correlates of outcomes with particular attention to associations between outcome and hospital and surgeon volume of THR.

Contact: Jeffrey N. Katz, MD, MS, Rheumatology Division, Brigham and Women's Hospital, 75 Francis Street, Boston, MA 02115.

Development and Implementation of a Total Joint Replacement Critical Pathway

The purpose of this project was to develop and implement a critical pathway for patients undergoing Total Joint Replacements within Nash Health Care Systems. Project goals were to: 1) decrease length of stay and costs, and simultaneously, 2) improve patient satisfaction and outcomes. An interdisciplinary team approach was used to develop the pathway involving physicians, nurses, rehabilitation staff and home health agency representatives.

A large part of the clinical pathway focuses on pre-operative education of the patient and/or family to provide an overview of what to expect regarding their hospital stay and their recovery and rehabilitation over time. A "Joint Prosthetic Device Fair" was conducted for surgeons to evaluate different prostheses and agree upon a vender to provide an appropriate joint for the best price. Goals to decrease LOS and costs have been achieved, I.E., LOS decreased from 10.4 to 4.2 days and costs were reduced 42 percent from $30,000 to $17,500.

The process to collect patient-based health status assessment information is underway. Decisions were made to: 1) use the SF-12 versus the SF-36, 2) have "pre-op nurses" collect this information over the phone before surgery, and 3) have data collectors in the Quality Improvement Department conduct a follow-up phone interview 3 months post surgery. Discussions are underway to determine if a 6 month follow-up interview will be conducted as well. Data collection is underway and the information system has been developed. Data management and analysis will begin in January, 1998.

Contact: Sandi Paige, RN, CCRN, Critical Pathway Coordinator, Nash Health Care Systems, 2460 Curtis Ellis Drive, Rocky Mount, NC 27804, (919) 443-8082.

Healthcare Research Systems Ltd. (HRS), is conducting patient perceived functional health status assessment for TJR patients on behalf of Alliant Health System (AHS), Louisville, KY, and Sarasota Memorial Hospital (SMH), Sarasota, FL. The instrument used is the SF-12 combined with an additional set of items which address issues of disease-specific functional assessment, risk-adjustment, patient satisfaction and infection control.

Alliant Health System surveys patients 3, 6 and 12 months post-operatively via telephone. The six month follow-up has recently been completed for the first cohort of patients. AHS and HRS are analyzing the data in an attempt to identify trends and relationships. Issues of risk-adjustment and relationships between satisfaction and self-perceived functional health status are of particular interest.

SMH surveys total joint patients pre-operatively and at 3, 6 and 12 months post-operatively via telephone. SMH is using the data collected pre-operatively in the patient education process. It is their intention to use data collected at that point to set realistic expectations with the patients about post-operative functionality. Of particular interest to SMH is the change in functioning over time, comparing pre-operative functioning to the final assessment at 12 months. Special attention is being paid to mental health information gleaned from the surveys to help identify patients who may need treatment for depression following surgery. It is the belief of SMH that mental health is an important component of physical improvement.

Contact: Melissa Kennedy, MHA, Healthcare Research Systems, 1650 Lakeshore Drive, Suite 300, Columbus, Ohio, 43204, (716) 244-0399.

Musculoskeletal Function Assessment: Long Form (MFA) and Short Form (SMFA) Versions

The MFA and SMFA were developed in response to the need for a health status instrument that could be used with a broad range of patients with musculoskeletal problems including fractures, soft tissue injuries, repetitive motion disorders and arthritis. The long version MFA has 100 items grouped into 10 categories including self-care, sleep, fine motor, mobility, housework, work, recreation, family relationships, cognition and emotional adjustment. The MFA takes about 15 minutes to complete. The MFA was developed and tested over three trials. Reliability was assessed by test-retest stability and internal consistency analyses. Content validity was based on a review of item selection procedures, expert opinion and the MFA's distribution of scores. Criterion validity was tested against physician ratings and standard clinical measures. Construct validity was demonstrated in an analysis of known groups, against existing health status measures, and in accordance with clinic hypotheses about the effect of musculoskeletal disorders on functioning. Using items from the MFA, the two part SMFA with 46 items was developed. It has been tested for reliability, validity and responsiveness with 420 patients from 18 community and university practices.

Contact: Julie Agel, A.T.C., Research Coordinator, Musculoskeletal Function Assessment, Harborview Medical Center, 325 9th Avenue, Box 359798, Seattle, WA 98104, telephone (206) 731-4113, Fax (206) 731-3227, Email: bagel@u.washington.edu.

The Impact of Preoperative Home Assessment and Physical Therapy on
Functional Outcomes of THA Patients

This study examined the impact of pre-operative home assessment and/or physical therapy on functional outcomes, length of stay, and total cost of total hip arthroplasty (THA) patients. Twenty-eight patients were randomly assigned to one of four groups: I. Standard clinical pathway only, II. Three pre-operative home based physical therapy sessions, III. One pre-operative home assessment visit by a registered nurse, and IV. Both pre-operative physical therapy and home assessment.

Functional status was measured using the MOS SF- 36 at the pre-operative medical office visit and six-weeks post-operatively. Post-operative home-based physical therapy assessments of hip strength, ability to transfer, inside ambulation, and stair climbing were made at admission to and discharge from this service. MOS SF-36 measures of bodily pain, physical functioning, and general health all indicated improvement. However, there was no significant difference across the 4 groups on these measures six weeks post-op.

In general it does not appear that pre-operative home care reduced length of stay or costs, or that it improved functional outcomes. Providing the same information using lower cost strategies maybe preferable. We are currently evaluating a class conducted by nursing, discharge planning, occupational therapy, and physical therapy which was introduced to address these issues. A video tape was also developed for THA patients and is distributed by the medical offices. Both of these educational strategies are aimed at providing the maximum amount of information in the most efficient and effective manner for the patient as well as the staff.

Contact: Sandra Gaynor, RN, DNSc, Director Nursing Development, Northwestern Memorial Hospital, 250 E. Superior, #1705, Chicago, IL 60611. phone: (312) 908-7756; fax: (312) 908-1741.

Fries JF, Spitz PW, Kraines RG and Holman HR (1980). Measure of patient outcome in arthritis. Arthritis and Rheumatism, 23:137-45.

Meenan RF, Gertman PM and Mason JH (1980). Measuring health status in arthritis: The arthritis impact measurement scales. Arthritis and Rheumatism, 23:146-52.

Meenan, RF and Mason JH (1990). AIMS2 Users' Guide. Boston, MA: Boston University School of Medicine, Department of Public Health.

Jette AM (1980). The Functional Status Index: Reliability of a chronic disease evaluation instrument. Arthives of Physical Medicine and Rehabilitation, 61:395-401.

Jette AM, Davies AR, Cleary PD et al. (1986). The Functional Status Questionnaire: Reliability and validity when used in primary care. Journal of General and Internal Medicine, 1:143-9.

Bellamy N, Buchanan WW, Goldsmith CH et al. (1988). Validation study of WOMAC: A health status instrument for measuring clinically important patient relevant outcomes to antirheumatic drug therapy in patients with osteoarthritis of the hip or knee. Journal of Rheumatology, 15:1833-40.

Tugwell P, Bombardier C, Buchanan, WW et al. (1987). The MACTAR Patient Preference Disability Questionnaire: an individualized functional priority approach for assessing improvement in physical disability in clinical trials in rheumatoid arthritis. Journal of Rheumatology, 14:446-451.

Martin J, Meltzer H and Elliot D (1988). The Prevalence of Disability Among Adults. OPCS Surveys of Disability in Great Britain, Report No. 1. London: HMSO.

Reisine ST, Goodenow C and Grady KE (1987). The impact of rheumatoid arthritis on the homemaker. Social Science and Medicine, 25:89-95.

O'Boyle CA, McGee H, Hickey A et al. (1992). Individual quality of life in patients undergoing hip replacement. Lancet, 339:1088-91.

Hip Replacement TyPE Specification, Liang MH, Health Outcomes Institute, 1990. For information contact Lisa Mills at (612) 858-9188.

Scott B, Brook RH, Lohr KN and Goldberg GA (1981). Conceptualisation and Measurement of Health for Adults. Vol. 6: Anemia. Santa Monica, CA: Rand Corporation.

Nicassio M, Wallston A, Callahan L et al. (1985). The measurement of helplessness in rheumatoid arthritis: The development of the Arthritis Helplessness Index. Journal of Rheumatology, 12:462-7.

EURIDISS (1990). European research on incapacitating diseases and social support. International Journal of Health Sciences, 1:217-28.

McDowell IW, Martini, CJM, Waugh W (1978). A method for self-assessment of disability before and after hip replacement operations. British Medical Journal, 2:857-859.

Bombardier C, Melfi CA, Paul J, Green R, Hawker G, Wright J, et al. Comparison of a generic and a disease-specific measure of pain and physical function after knee replacement surgery. Medical Care 1995;33(4):AS131-AS144.

Hawker G, Melfi C, Paul J, Green R, and Bombardier C. Comparison of a generic (SF-36) and a disease specific (WOMAC) instrument in the measurement of outcomes after knee replacement surgery. Journal of Rheumatology 1995;22:1193-1196.

Kantz ME, Harris WJ, Levitsky K, Ware JE, and Davies AR. Methods for assessing condition-specific and generic functional status outcomes after total knee replacement. Medical Care 1992;30(Suppl 5):MS240-MS252.

Stucki G, Liang MH, Phillips C, and Katz JN. The Short Form-36 is preferable to the SIP as a generic health status measure in patients undergoing elective total hip arthroplasty. Arthritis Care and Research 1995;8(3):174-181.

Williams JI, Young N, and Arshinoff R. Comparing a disease-specifi functional status measurement in assessing c measure and a the waiting times and outcomes of hip and knee replacements. Quality of Life Research 1994;3:99.

Young N, Williams JI, Wright JG, and Piexto G. Rating difficulty of functioning in patients undergoing hip or knee replacements: comparing the WOMAC and Physical Functioning (SF-36) Scales. Quality of Life Research 1994;3:100.

Katz JN, Larson MG, Phillips CB, Fossel AH, and Liang MH (1992). Comparative Measurement Sensitivity of Short and Longer Health Status Instruments. Medical Care, 30, 917-925. (Also SF-36)

Liang MH, Fossel AH, and Larson MG (1990). Comparisons of Five Health Status Instruments for Orthopedic Evaluation. Medical Care, 28, 632-642.

Weinberger M, Samsa GP, Tierney WM, Belyea MJ, and Hiner SL (1992). Generic versus Disease Specific Health Status Measures: Comparing the Sickness Impact Profile and the Arthritis Impact Measurement Scales. Journal of Rheumatology, 19, 543-546.

Calfas CJ, Kaplan RM, Ingrum RE (1992). One-year evaluation of cognitive-behavioral intervention in osteoarthritis. Arthritis Care and Research, 5, 202-209.

Kaplan RM, Kozin F, Anderson JP (1988). Measuring quality of life in arthritis patient. Quality of Life in Cardiovascular Diseases, 4, 131-139.

Kaplan RM, Anderson JP, Wu A, Mathews WC, Kozin FL, and Orenstein D (1989). The quality of well-being scale: Applications in AIDS, cystic fibrosis, and arthritis. Medical Care, 27 Suppl 3, S27-S43.

Health state utilities in knee replacement surgery: the development and evaluation of McKnee.

Bennett KJ, Torrance GW, Moran LA, Smith F; Goldsmith CH. Journal of Rheumatology. 1997 Sep;24(9):1796-805.

The purpose of this study was to: 1) develop McKnee, a classification system and direct utility measure for health states associated with knee replacement (KR) surgery, and 2) apply McKnee in a before-after study of KR surgery to: (a) gain experience with McKnee in an elderly population; (b) confirm the practicality and usefulness of the McKnee system; (c) assess self-health utility one week before and 3 months after surgery; (d) evaluate the stability of 3 clinical marker health states describing mild, moderate, and severe knee disability; and (e) compare self-health utility scores with Short Form 36 (SF-36). Among the study findings: 1) McKnee was feasible and acceptable in the older patient group studied, 2) no change was observed in the self-health utility at 3 months postsurgery; 3) on the SF-36, only the change scores for pain and health transition were statistically significant, and 4) the clinical marker mean utility scores were stable between the baseline and 3 months assessment, but the intraclass correlation coefficients for individual scores were low. The authors concluded that McKnee can be used to obtain and interpret the knee disability utility scores in cost-utility studies and medical decision-making models. It is a practical and useful method for classifying knee disability health states.

Quality of life after knee arthroplasty. A randomized study of 3 designs in 42 patients, compared after 4 years.

Hilding MB, Backbro B, Ryd L. Acta Orthopaedic Scandanavia. 1997 Apr;68(2):156-60.

We assessed different yardsticks for outcomes 3-5 years after surgery in a prospective, randomized study of 42 patients, where 3 designs of cementless knee prostheses were used. The prognosis with regard to loosening, previously obtained by radiostereometry after 2 years of follow-up, was utilized. Patients with a prognosis of stable implant fixation (two thirds) were compared with those where loosening was predicted (one third). Hospital for Special Surgery score and Visual Analogue Scales regarding pain at rest, "first step" pain, pain during activity and global function, showed consistent postoperative improvements, but no differences between the design and prognosis groups were found. Quality of life assessment by the Nottingham Health Profile questionnaire showed that the poor prognosis group had increased pain and significant disturbances of sleep and emotions, as well as difficulty in enjoying hobbies and holiday activities. No differences were found between the design groups. Altogether, the patients showed profiles comparable to a healthy reference group. We conclude that the Nottingham Health Profile is a sensitive, relevant and simple measure of outcome after knee arthroplasty.

Williams JI, Llewellyn Thomas H, Arshinoff R, Young N, Naylor CD. Journal of Evaluation of Clinical Practice.1997 Feb;3(1):59-68.

The objectives of this study were to assess the impact of major joint replacements in reducing pain and disability and to describe the burden of pain and disability that could be avoided by ordering the queues with respect to severity of disease. A secondary goal was to compare the uses of a general health status measure, the Short Form Health Survey (SF-36), and a disease-specific measure, the Western Ontario McMaster Osteoarthritis Index (WOMAC), for accomplishing the objectives. The results are based on interviews with 209 patients before and after surgery. Only 15.9% of the patients had surgery within 3 months' waiting time, 19.2% waited 4-6 months, 30.7% waited 7-9 months, and the remaining 34.1% waited a year or more. The waiting times were unrelated to the severity of pain or disability reported in the initial interview.

Following surgery, there were large reductions in the WOMAC scores for pain, stiffness and difficulty in functioning. The SF-36 showed substantial improvements in relief from pain and in physical functioning, and reductions in role limitation due to physical problems, but not for scores related to mental health. The WOMAC scores were more responsive to the benefits of surgery than the SF-36 scores. Queuing systems keyed on burden of symptoms could reduce the burden of pain and disability suffered by patients awaiting surgery. The improvements from hip and knee replacements suggest that equitable access for these procedures should be a priority in Ontario.

Franzen H, Johnsson R, Nilsson LT. Journal of Arthroplasty. 1997 Jan;12(1):21-4.

The functional results in terms of quality of life 10 to 20 years after cemented total hip arthroplasty due to primary arthrosis were evaluated in 187 patients with nonrevised hips by use of the Nottingham Health Profile questionnaire. Function was impaired compared with age- and sex-matched random control subjects.

Functional outcome and patient satisfaction in total knee patients over the age of 75.

Anderson JG, Wixson RL, Tsai D, Stulberg SD, Chang RW. Journal of Arthroplasty. 1996 Oct; 11(7):831-40

Seventy-four patients, age 75 or older, who had undergone 98 primary total knee arthroplasties were evaluated in a retrospective cohort study, with validated questionnaires that assessed self-reported pain, physical function, mental health, and satisfaction. Average follow-up period was 34 months (range, 12-67 months). Overall, 90.8% reported improvement, 88.8% were satisfied with the results of surgery, and 91.8% felt they had made the right decision. Dissatisfaction with the results correlated with poorer mental health scores, decreased physical function, and increased bodily pain scores (P < .05). Satisfaction was correlated with better pain scores on the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) and SF-36 (P < .05) but not with Hospital for Special Surgery scores (P = .328). Poor surgical results leading to revision surgery (5%) were associated with preoperative deformity greater than 20 degrees. Based on this patient-assessed outcome analysis, total knee arthroplasty is a worthwhile and beneficial procedure in the elderly.

In a prospective trial we studied 176 consecutive patients having a primary total hip arthroplasty to compare the quality of life before and after operation in non-obese and obese groups. We used a modified Harris hip score and the Rosser Index Matrix to generate these scores, and found that the median preoperative quality of life score for both groups was similar. During review, both groups showed a corresponding marked improvement in quality of life scores at one and three years. There was no statistically significant difference in the improvement in scores between the non-obese and obese groups. It appears that relative body-weight on its own does not influence the benefit derived from primary total hip arthroplasty.

Aarons H, Hall G, Hughes S, Salmon P. Journal of Bone & Joint Surgery. 1996 Jul;78(4):555-8

There are many studies of long-term recovery from major joint arthroplasty, but little is known about the first days and weeks after operation. The purpose of this study was to measure function, emotional state and life evaluation before arthroplasty and at seven and 50 days after in a consecutive series of 40 hip and 23 knee replacements. Pain was relieved significantly at seven days after hip arthroplasty and even more at 50 days. In knee patients, pain relief was modest and was not apparent until 50 days. Functional ability was much improved by 50 days in hip patients, but hardly changed in knee patients. Positive mood and life satisfaction did not improve in either group.

The Nicolas Andry award: comparative results of cemented and cementless total hip arthroplasty.

Rorabeck CH, Bourne RB, Mulliken BD, Nayak N, Laupacis A, Tugwell P, Feeney D. Clinical Orthopaedics. 1996 Apr;(325):330-44.

A prospective randomized clinical trial comparing cemented to cementless total hip arthroplasty has been ongoing since 1987. All parameters assessed showed that there was no statistically significant difference in any of the health-related quality of life measures used. The improvement was dramatic and sustained. No components (cemented or cementless) have been revised to date (average, 4.8 years). Total hip replacement gives comparable clinical results when health related quality of life measures are analyzed; however, metal backed cemented acetabular components have a high rate of radiographic failure. Short-term stability of a non-cemented porous coated titanium stem, however, gave excellent clinical and radiographic results.

Quality of life and functional ability in hip and knee replacements: a prospective study.

Rissanen P, Aro S, Sintonen H, Slatis P, Paavolainen P. Quality of Life Research. 1996 Feb;5(1):56-64.

The impact of hip (THA) and knee arthroplasty (TKA) on patients' health-related quality of life (HRQOL), physical ability and functioning was assessed in a two year follow-up study of 276 hip and 176 knee patients. The eligibility criteria were a diagnosis of primary arthrosis, a primary operation, and total joint arthroplasty. Patients were interviewed by questionnaire prior to the operation and 6, 12 and 24 months after the surgery. Subjective health outcomes were assessed with the Nottingham Health Profile and the 15D, a fifteen dimensional HRQOL measure. Patients' physical ability was assessed using measures of activities of daily living, and of physical mobility. Patient related outcome variations were analyzed by regression models. Major improvements were observed for pain, sleep and physical mobility. On average, in most of the quality of life dimensions the patients attained a similar quality of life as the comparable general population and only 4.7% of hip and 9.7% of knee patients had a worse HRQOL score at all three post-operative measurements than at baseline. Naturally, those with the poorest HRQOL pre-operatively gained most from the operation. High age did not lessen HRQOL gains from THA, but in TKA the oldest patients gained least in terms of 15D scores. Hip, but not knee patients with a long education tended to have greater improvements in quality of life and functional ability.

Norman-Taylor FH, Palmer CR, Villar RN. Jounral of Bone & Joint Surgery. 1996 Jan;78(1):74-7.

Total hip and total knee arthroplasty both provide a considerable improvement in quality of life, but there is no evidence to suggest that one is more successful than the other. We studied 72 patients in a prospective trial before and after total hip or total knee replacement. We recorded scores for disability and distress derived from the Harris hip score and the British Orthopaedic Association knee assessment score, and used them to generate quality of life (QoL) scores using the Rosser Index Matrix immediately before and at one year after surgery. The patients awaiting knee replacement had significantly lower QoL scores than those awaiting hip replacement (p = 0.01). The QoL scores at one year post surgery were high and almost identical for both groups (p = 0.46). Further analysis showed that gender and weight were not significant predictors of improvement of QoL scores, but age (p = 0.03) and whether the hip or knee was replaced (p = 0.006) were significant factors.

Editor's Note:

This article was written by Arnold Milstein, MD, MPH. It was based on and summarizes the main points of Dr. Milstein's and Dr. Jim Franklin's presentation given at the Trust's State-of-the-Art Health Outcomes Conference in May, 1997.

Dr. Milstein is Medical Director of the Pacific Business Group on Health and a Managing Director at William M. Mercer Inc. Dr. Franklin directs benefits at Stanford University and, as a prominent member of PBGH, provides stewardship for Physician Group Value Check.

In 1994, America's largest healthcare purchasers' coalition, the Pacific Business Group on Health, agreed with a committee of its most utilized provider groups to use a longitudinal patient survey to measure functional status across all patients, as a means of comparing quality of care among California physician groups. The providers were seeking relief from what they experienced as a market which primarily rewarded low price. For them, the study offered a path for differentiating themselves on the basis of quality.

PBGH employers were also seeking to attach more significance to quality in relationships between providers and health plans, as well as between providers and their patients. Other PBGH goals included: (1) focusing greater physician attention on patient-perceived functional status; (2) adding a comprehensive index of outcome to available comparative quality measures which primarily addressed clinical processes associated with single conditions or treatments; and (3) creating publicly available quality comparisons at the physician group level. Comparison at the physician group level was felt to be especially important, since most California health plans use and delegate most quality management to overlapping physician groups. This makes exclusive reliance on plan-level comparisons of quality of care highly problematic.

This confluence of purchaser and provider interests was joined by PBGH's HMOs, which recognized that their current methods of assessing the quality of their physician group contractors were duplicative and did not address purchaser aspirations for broad longitudinal measures of quality based on patient-rated outcomes.

After consultation with Dr. Bill Rogers and Dr. John Ware from the Health Institute at New England Medical Center in Boston, PBGH and its provider committee selected the SF-12 as their core survey instrument, based on its scientific pedigree, reliable capture of SF-36 physical and mental component summary scores, and likelihood of higher response rates with a shorter survey than with the SF-36.

In pursuing a project of this complexity, PBGH utilized a strategy that rested on five key principles. First, PBGH sought the widest possible partnership with physician groups. It expanded pre-existing links with many larger physician groups through an alliance with The Medical Quality Commission (TMQC). TMQC was, until 1997, the quality assurance affiliate of the Unified Medical Group Association (UMGA), the largest integrated medical group organization in the West, now operating as the American Medical Group Association (AMGA). PBGH also allied with another large physician group organization, the IPA Association of California (now NIPAC). Second, PBGH championed public disclosure of results, since rewarding higher quality with more market share can be facilitated by sharing comparative quality information with patients. Third, PBGH adopted a bias for action and progress in comparative measurement, recognizing that imperfections in methodology and its implementation were inevitable and that absence of comparative quality information carried a large, if widely unrecognized, ongoing cost.

Additionally, PBGH realized that capturing differences in effectiveness of medical groups in their longitudinal management of functional status in non-elderly populations would require large sample sizes. Dr. Bill Rogers estimated that detecting significant differences among physician groups for populations of 50-70 years old over a two-year interval via the SF-12 would require an initial sample of over 600 patients per physician group. Accordingly, PBGH sought to augment its own substantial employer-funded Quality Improvement Fund with additional financial support, which was forthcoming from several progressive sources such as the Kaiser Family Foundation via Dr. Mark Smith and pharmaceutical companies, led by Schering-Plough. Finally, wide collaboration with health service researchers was pursued, with particular focus on individuals with prior experience in large scale use of longitudinal patient survey to judge quality. Drs. Bill Rogers, John Ware, Ron Hays (RAND) and Anita Stewart (UCSF) were particularly helpful.

In 1995, the Medical Quality Commission joined PBGH in sponsoring the study which was named Physician Group Value Check. This was a first-time collaboration between the PBGH and TMQC that also drew from previous activities each group had pursued. PBGH had conducted an annual survey of enrollees in health plans offered by its employer members and was considering extension to the larger medical groups serving its employees. TMQC had conducted its own survey of medical group enrollees which focused on patient satisfaction. The PBGH-TMQC initial collaboration was marked by substantial concurrence on measurement objectives.

Physician Group Value Check is a comparative assessment of physician group performance over time. The time-one survey occurred in the Fall of 1996; the time-two follow-up survey with the same patient cohort is scheduled to occur in the Fall of 1998. Fifty-eight multispecialty medical groups and IPAs (primarily based in California) participated. The time-one cohort for each participating physician group included 1,000 patients who received physician services in the prior twelve months. All patients included in the study were HMO enrollees. A comparison group of patients who were not enrolled in HMOs and whose care was not under the direction of an organized physician group was constructed with the help of an indemnity plan insurer.

CLINICAL PEARLS

Predicted Highest Yield Strategies to Optimize Physical
and Mental Funtioning Among 50-70 Year Old HMO Enrollees Over a 24 Month Time Period

Broad Interventions with Patients

Aggressively promote endurance walking and strength training via a case management paradigm

Risk-Specific Interventions with Patients

Aggressive identification and case management of individuals who smoke, are sedentary, lack social support or have undergone recent major loss
Identify high risk individuals via periodic survey; then assure physician counseling for enrollees engaging in high risk behaviors and self-management education for the chronically ill
Implement substantial programs aimed at regular exercise, smoking cessation, group weight loss, and, for home-bound disabled, internet-based social networks

Broad Interventions with Clinicians

Engage physicians in fundamentally re-focusing their primary effort from treating diagnoses to treating patient-defined functional impairments
Use periodic enrollee assessment of health status and clinician helpfulness to assure that pre-determined comprehensive clinical protocols are being applied seamlessly at every level of patient functional status

Risk-Specific Interventions with Clinicians

Administer the SF-12 to all patients, identify the diagnoses capturing the greatest number of low scores, and systematically re-engineer clinical are for these diagnoses

Patient information was collected with one questionnaire to measure physical and mental components of health status, satisfaction, receipt of preventive services, and receipt of selected services for patients who were aware they suffered from high blood cholesterol or hypertension. The survey was administered through the mail with phone follow-up. Physical and mental components of health status will be measured using the SF-12. Survey items needed to risk-adjust change in health status over the 24-month period will be drawn from questions recommended by researchers at the Health Institute, who will join RAND in data analysis. The preventive services and satisfaction survey items are derived from the Health Institute questions that PBGH annually asks of health plan enrollees, as well as from a prior medical group-specific questionnaire co-developed by TMQC and RAND. The survey questions pertaining to receipt of services targeted to patients with high blood cholesterol and hypertension were drawn from RAND publications from the Health Insurance Experiment.

The study will also collect from each participating physician group information on various key aspects of their clinical operations and reimbursement relationships with HMOs and their member physicians. The relationships between these structural and operational variables and the outcomes being measured with the patient survey will be quantitatively evaluated. The evaluation of risk-adjusted change in population health status will focus on the 50-70 year old cadre, which will be oversampled to support detection of differences between physician groups. Seven hundred of the 1,000 enrollees surveyed in each physician group will be 50 to 70 years old. The study is being coordinated by PBGH's Director of Research, Dr. Cheryl Damberg, and the author, in close partnership with Dr. Alan Zwerner and Lori Bloomfield of TMQC.

"You get what you measure" is a frequent observation emerging from current American efforts to advance quality accountability in health care. Centering Physician Group Value Check on broad population indices of longitudinal physical and mental functioning was a purposeful decision rooted in an employer's perspective on how clinical services translate into customer value. PBGH employers recognize that successful management of patient-rated functional status is what generates gain for employers via improved work attendance and on-the-job employee productivity. They also recognize that maximum reduction of the impact of illness and health risk on an individual's ability to function in life comprises the essential purpose of employees and their dependents in seeking health care.

In contrast, current American medical practice is commonly focused on the narrower objectives of making a correct diagnosis and delivering those treatments such as prescription-writing which require a medical license to initiate. As Stanford's Dr. Harold Holman and other analysts of American healthcare have observed, this focus is better suited to the inpatient care of acute illness than to the ambulatory care of our predominant national health problem of chronic illness. Of particular relevance is recognition that health status is primarily governed by health behavior and environment. Today, very few American delivery systems are engineered to substantially affect those two primary drivers. For example, prior PBGH surveys of health plan enrollees show less than 30% compliance with U.S. Preventive Service Task Force recommendations for every patient age group on counseling of patients by clinicians to mitigate major behavioral and environmental health risks.

To help physician groups participating in Physician Group Value Check to close this gap between customer aspirations for functional status and current delivery system emphasis, PBGH sought advice from nationally distinguished physicians who have been active in improving clinical attention to patient-rated functional status. They were asked individually to summarize for participating physician groups a strategy likely to generate the greatest gain in physical and mental functioning over 24 months in a 50 to 70 year old cohort of HMO enrollees participating in Physician Group Value Check. Responses were obtained from Dr. Donald Berwick (Institute for Healthcare Improvement), Dr. Harold Holman (Stanford University), Dr. Stephen Jencks (HCFA), Dr. Jeffrey Kang (HCFA), Dr. Haya Rubin (Johns Hopkins), Dr. Barbara Starfield (Johns Hopkins), Dr. Ed Wagner (Group Health Cooperative of Puget Sound) and Dr. John Wasson (Dartmouth). Their advice is summarized in the Table of "Clinical Pearls." These pearls were then shared with participating medical groups to make more tangible the concept of clinician management of global functional status.

The Table of Clinical Pearls comprises an early, provocative glimpse at the nature of the armamentarium of population health management. Though there are discernible common threads among the strategies, their diversity confirms that the science of medical management of populations is not more advanced than the science of clinician care for individual patients. The pearls fell into four broad categories, varying in scope (pertaining to all patients versus patients with particular risks/conditions) and in mode (interventions aimed at changing patient behavior versus changing clinician behavior). As with the care of individual patients, management of populations will gradually expand its evidence base, highest-yield interventions will be more easily discerned, and variations among delivery systems gradually diminish. In the meanwhile, groups participating in Physician Group Value Check can use these recommendations from among today's most experienced and innovative change agents to shape their strategies for success in managing functional status across large populations.

How quickly will American delivery systems make the transition to a focus on value-added, as defined by their patients' perspective on their ability to function? Momentum generated by PBGH's initiative will be reinforced at the national level among HMOs by a "sister" quality measure, Health of Seniors, within the NCQA HEDIS 3.0 measure set. While the pace of change is not predictable, Physician Group Value Check will help stimulate forward movement among many of America's largest physician groups by recognizing and rewarding provider success, while creating a balance to cost-focused appraisals of their comparative value.

The Health Care Financing Administration (HCFA) is demonstrating notable leadership in the still formative movement to integrate the use of patient-based health outcomes into US health care systems. As discussed in the July and October 1997 issues of the Trust's Monitor, HCFA has spearheaded the development and implementation of systems to incorporate health status and outcomes measures of health-related quality of life in managed care organizations serving the Medicare population and in nursing homes. The Health of Seniors (HOS) Outcome Measure initiative and the Minimum Data Set for Nursing Home Assessment and Care Screening (MDS) program, respectively, collect data to measure patient-based health-related quality of life in managed care and nursing home populations. The MDS was implemented in the early 1990's and the HOS initiative is in the early implementation phase.

The field of health outcomes has advanced tremendously with these two initiatives, e.g., infrastructures that take years to build have been created and/or are under development, psychometrically sound data are being generated, and patient-based outcome information is being used to improve patient care. The purpose of this article is to introduce a third HCFA-sponsored, outcomes-based system improvement, the Outcomes and ASsessment Information Set (OASIS), an outcome-based quality initiative intended to continuously improve the quality of formal home care services.

The increasing need for long-term care in the United States and its link to the growing old and oldest-old populations have been widely documented. According to the U.S. Census Bureau, individuals 65 years of age and older make up almost 13 percent of the US population and are expected to comprise over 20 percent by the year 2030. Moreover, persons 85 years of age and older will continue to make up an increasing proportion of the elder population. This segment of the population which increased 274 percent from 1960 to 1994, has the highest rates of chronic disease and functional disability of any group, as well as high per capita health care costs. In addition to the increase in total demand for long-term care services, home care and nursing home providers have noted a need for: 1) a more diverse array of services, and 2) clinical services to meet the needs of a sicker population.¹

Home health care is the fastest growing segment of the long-term care market. In the past six years, Medicare spending for home health care has ballooned from $2 billion to $14 billion.2 The growth rate of home health care is faster than all other types of Medicare-covered health care with projected expenditures expected to increase almost 11 percent per year through 1999. Furthermore, other public funding sources, e.g., Medicaid, are experiencing a similar phenomenon for home and community-based services.1 As in all health care delivery sectors, rising costs have lead to efforts to understand how much health is being purchased with health care dollars.

OASIS is an outcome-based quality initiative intended to continuously improve the quality of formal home care services. Although OASIS collects patient-based health-related quality of information, it does not collect it directly from the patient. Rather, a clinician familiar with the patient completes the instrument on his/her behalf. OASIS or OASIS-B, the second and currently operative version of the tool, includes 79 items collecting information in the following 13 domains:

Most of the data collected with OASIS - B are health status measures or measures that when administered over time will generate information about change in patients' status or outcomes. Peter Shaughnessy, PhD, director of the Center for Health Sciences and Policy Research at the University of Colorado and the principal investigator of projects to develop and evaluate OASIS under contract with HCFA provided the perspective that OASIS is intended to answer the question "what are we getting?" for the $14 billion we are spending on home health care.²

OASIS had its genesis in a five-year research program funded by HCFA and the Robert Wood Johnson (RWJ) Foundation to develop outcome measures for home care. OASIS-A, (the first iteration of the instrument) was developed and tested in a national research program and then used operationally in two demonstration projects. A national demonstration project was funded by HCFA and an analogous project, funded by RWJ, was conducted in New York State. Both of the demonstration projects were administered by the University of Colorado Research Center. The experience of the 50 national demonstration agencies and the 22 New York State demonstration agencies using OASIS were applied in the modest set of revisions that resulted in OASIS-B. Reliability testing, programmatic applications, and provider input to improve OASIS will continue with a view toward improving the data set over time. Nonetheless, OASIS is now regarded as a stable data set that can be used in the context of patient assessment and outcome monitoring.³

Deliberations took place about whether it would be wise for the Research Center to release OASIS-A in August 1995, since, within a minimum of 18 months, it would be followed by OASIS-B. In keeping with its philosophy of establishing a partnership with the home health industry, HCFA staff determined it was appropriate to provide the industry with each version of OASIS as they become available. In addition to keeping with the philosophy of open communication and information sharing, providing home health agencies with OASIS-A was seen as a way of: (1) preparing home health agencies for the new Medicare approach to data collection required for purposes of monitoring outcomes, (2) assisting providers in collecting more precise data for purposes of assessment, (3) allowing providers to begin to fit the OASIS data items into their agency's tailored approach to data collection (while OASIS will be a standardized approach to collecting information, home health agencies will fit it within a more comprehensive approach to data collection that is unique to each agency, e.g., the collection of more process-oriented information), and (4) assisting the industry and Medicare in continuing to use an important data set before it is mandated for use. ³

The current status of OASIS-B is that HCFA has distributed proposed rules or "Conditions of Participation" for Medicare and Medicaid-certified home health agencies with regard to the use of OASIS. This information was distributed in the March 10, 1997 issue of the Federal Register.⁴ In it, HCFA announced the specific changes home health agencies will need to comply with in order to participate in the Medicare and Medicaid programs once these regulations are formalized. With this notice of proposed changes, HCFA entered a period during which input was solicited from home health agencies, consumers and any other interested constituency. Currently, HCFA is reviewing and integrating the input received from the public and anticipates the release of the final Conditions of Participation that will require the use of OASIS in the second half of 1998.

1. Peter W. Shaughnessy, Robert E. Schlenker, Kathryn Crisler, Angela Arnold, Martha C. Powell, and Hames M. Beaudry. "Home Care: Moving Forward with Continuous Quality Improvement." Journal of Aging and Social Policy, 1996. 7:149-167.

2. "Clinton aims to order all home care agencies to collect outcomes." Medical Guidelines & Outcomes Research. March 20, 1997. 8:1-2.

3. Peter W. Shaughessy, Kathryn S. Crisler, and Robert E. Schlenker. "Medicare's OASIS: Standardized Outcome and Assessment Information Set for Home Health Care - OASIS-B." Distributed by the National Association for Home Care. March, 1997.

4. "Medicare and Medicaid Program: Revision of Conditions of Participation for Home Health Agencies and Use of Outcome Assessment Information Set (OASIS); Proposed Rules" in the Federal Register, March 10, 1997.

Editor's Note:

This article, written by Deb Lowry, summarizes her presentation given at the Trust's third State-of-the-Art Health Outcomes Conference in May, 1997.

Deb Lowry is the Director of Market Performance Appraisal for Kaiser Permanente and has been based on the national Program Office since 1990. Kaiser Permanente, a not-for-profit managed health care organizations with a 51 year history; it serves over eight million members in 18 states.

The purpose of this article is to discuss the impact of national initiatives in performance assessment on managed care organizations. The article will cover four broad themes including: 1) challenges that initiatives in performance assessment pose to managed care organizations (MCO), 2) the purposes for which purchasers use performance assessment information, 3) HEDIS, and 4) Kaiser Permanente's strategies for dealing with challenges.

As the quality and value of health care take a more prominent place on the national agenda, there are an increasing number of initiatives devoted to a wide range of performance assessment activities. Gathering and organizing data, and generating information reports to respond to a growing number of mandates pose significant challenges to MCO. This section of the article presents eight of those challenges:

The first challenge to MCO, the requirement to measure health plan performance for the "enrolled population", may seem straightforward. But that perspective is based on an assumption that plans' enrolled populations are static, that they remain the same from the start to the finish of a reporting period. In fact, a health plan's enrolled population is always changing. Members disenroll at an average annual rate of 10 to 15 percent and, as they do, new members enroll to take their place. The difficulties in dealing with these occurrences relate to the complications that different time windows (within which enrolled populations fall) present to virtually all of the tasks inherent to measurement, IE, data collection, programming, analysis and interpretation of data.

The next two challenges are interrelated. The first is the aging of the population. The baby boom generation, steadily advancing into middle age, represents a large and increasing segment of the health care consumer population. Simultaneously, we have entered an era of consumerism in which people seeking health care are expecting and demanding improved service in many ways, e.g., more convenient appointment times such as evening and week-end office hours. Just as the health care establishment and its many players are focusing on improving health outcomes and satisfaction (and offering comprehensive services at the lowest possible cost) the baby boomers, who tend to drive the consumer satisfaction movement, are entering an age in which they are making up an increasing proportion of the health care user population. The fact that consumers are expecting more from their physicians and other providers is a good thing. But it does present challenges to MCO, e.g., assertive consumers balk at the idea of not spending enough time with their physicians at a time when providers are under pressure to keep patient visits short and "cost effective". Situations such as this are requiring managers and clinicians in MCO to think creatively and to develop new strategies to achieve objectives that sometimes seem at odds with one another.

Another set of challenges relates to the introduction of new populations into managed care settings. Most notable recently, Medicare and Medicaid beneficiaries are entering managed health care settings in large numbers. Both populations were among the fastest growing segments of Kaiser Permanente's enrolled population in 1997. However, each population tends to be characterized by different attributes that result in unique service delivery and management challenges.

The fifth challenge facing the managed care industry and those involved in the movement to continuously improve care is the issue of patient confidentiality and the protection of members' privacy. Individuals seeking health care depend on their doctors to protect their privacy and yet, increasingly, MCO are required to share more and more information about patients' status and the care they receive with an increasing number of organizations involved in some aspect of the oversight of health care.

Developing cost effective methods to respond to both public and private sector demands for information is another challenge with which the managed care industry is dealing. General Motors has quoted the cost of health care for their employees relative to the costs of a new car. They report that the health care for their workforce amounts to $1,200 per car, a figure that demonstrates why it is in all of our interest to do what we can to maximize the value of each health care dollar.

One of the specific challenges that Kaiser Permanente faces (having been around for 51 years) is integrating computers into infrastructures that were created before computers existed. The collection of better data and obtaining ready access to it requires placing computers in physicians' workstations. Many of the offices at Kaiser Permanente required remodeling and rewiring to accommodate today's technology. Creating the technological infrastructures needed to access, manipulate and manage data is, in one way or another, a challenge with which all organizations in health care are dealing.

The final challenge to MCO discussed in this article relates to the "information age" in which we live. Public and private organizations to whom provider are accountable (and others in all sectors of society) expect and demand access to a tremendous amount of information. Providers are required to respond to many requests for data and information from: 1) private insurers and their benefit consultants, and 2) public purchasers, e.g., HCFA, state Medicaid agencies, the Office of Personnel Management (OPM), other state organizations involved in the delivery of care, and public retirement systems. Although many of these requests are similar, they are often redundant and duplicative, and responding to them draws on many resources. Generally, purchasers' requests are for data and information in the following areas: descriptive information about plans' administrative policies and procedures, management and finances, quality, utilization, member satisfaction, and access to service.

When MCO charge that purchasers make too many requests for information, purchasers say they need different data in different forms for different reasons. They offer a list of purposes for which they use information including:

Chief among purchasers' uses for the data MCO provide is to make plan-to-plan comparisons and to give that information to their employees to make decisions about which plan to choose. One example of this approach comes from Xerox which, for approximately three years, has provided side-by-side comparisons of health plans to their employees at the time of the annual open enrollment season. If plans do not provide the comparative metrics, Xerox employees want to know why the plans continue to be included among their health care choices.

Some purchasers use the data MCO provide to negotiate fees with them and /or as the basis for narrowing the selection of health plans they offer. Again, Xerox serves as an example. For a period of four to five years, Xerox has been using performance data to encourage employees to consider managed care options. During this time, they have seen a shift from 35 to 70 percent of their employees enrolling in managed care plans.

GTE offers another example of how purchasers use provider data to compare health plans and influence the marketplace. GTE has used provider data to contrast fee-for-service versus managed care plans in order to: 1) examine and understand geographic differences in practice and utilization patterns, and 2) keep health care costs down. In this context, GTE has sponsored carrier and provider seminars in many of their markets with the aims of reducing variation among its providers, improving health care quality and reducing costs. The thinking behind GTE's seminars is that the quality of health care of a community will improve if practice variation is diminished with concomitant increase in effectiveness, and that these improvements will, in turn, lead to a decrease in costs.

Another way that purchasers use performance data is to risk adjust premiums, a process that involves sophisticated analysis and data modeling. (See the article by Mark Hornbrook in the July 1997 issue of the Monitor and visit the Members' Room on the Trust's web site for a transcript of an expert conference call with Lisa Iezzoni.)

The Health Plan Employer Data and Information Set (HEDIS) is a set of standardized performance measures designed to provide purchasers and consumers with information to assist them in making reliable and meaningful comparisons of managed care plans. HEDIS is the most widely used system of performance measurement in managed care and therefore, is an important case study with which we can examine the impact of performance assessment initiatives on managed care. HEDIS began as a collaborative effort between purchasers of health care benefits and MCO with the support of the National Committee on Quality Assurance (NCQA). In addition to the original participants, the collaboration has grown to include scientists, consumers, public purchasers and labor unions. The many players involved with HEDIS demonstrate how much is at stake and how many stakeholders there are in the area of health care performance assessment.

In 1997, version 3.0, the third iteration of HEDIS, was implemented. It represents a significant step forward with regard to the benefits it delivers to the field of health care improvement. However, when compared with HEDIS 2.5, the previous version, there are also more costs associated with HEDIS 3.0. HEDIS 3.0 involves more measures (75 versus 60), more populations, more hybrid approaches to the development of measures (combining data from administrative data sets with data from medical records), more programming, more medical record reviews, more staff to do the work, more data to interpret, more to understand and more results to improve over time.

A specific note regarding HEDIS 3.0 and its incorporation of patient-based outcomes measures follows. HEDIS 3.0 incorporates the Health of Seniors (HOS) Outcome Measure, operationalized with the SF-36. The HOS measure represents a major step forward in national performance assessment initiatives which has come about as a result of HCFA, NCQA and many other organizations who participated in a committee devoted to the development of the HOS measure. Prior to the introduction of the HOS measure, the measures included in the HEDIS data set were all process measures. (For a full discussion of the Health of Seniors measure, see the July 1997 issue of the Monitor.)

The results from HEDIS 3.0 and its earlier iterations are not comparable because of the significant changes in how the measures and reporting parameters are defined. Kaiser Permanente and other MCO have decided not to compare HEDIS 3.0 data with data from previous years. Instead, they will consider 1997 as the baseline year for HEDIS 3.0 reporting in the future. Health plans collaborated with the goal of finding a way to make valid comparisons between versions 2.5 and 3.0 but it was impossible to achieve in an affordable way. This is unfortunate because it prevents providers from looking at trend data from the three years that HEDIS data have been collected. However, because HEDIS is young and significant improvements have been made to it over time, this limitation is neither surprising nor unusual. Effectively, health plans have opted to trade-off short term comparisons for better quality data in the long run. Given this situation, it is critical that HEDIS data are analyzed, interpreted, reported and used with discretion. For example, comparing measures from 2.5 and 3.0 of childhood immunization or cervical cancer screening will likely generate an impression that performance has declined which may, in fact, be an artifact of the changes in measurement definition.

At the end of the first HEDIS reporting cycle, Kaiser Permanente staff realized their workforce was not being informed of the results of the assessment of their performance. In response to this communication gap, Kaiser Permanente created a publication called "How We Measure Up", a report to Kaiser Permanente's workforce with internal and external comparisons of their performance on HEDIS data and other performance measures. Other performance measures, e.g., satisfaction data, are also being used to incent executives, employees and physicians to improve service. Some performance measures have even found their way in to labor union contracts at Kaiser Permanente, e.g., incentives for achieving certain levels of member satisfaction with service.

Kaiser Permanente has been working to develop strategies to respond to the plethora of requests from purchasers for most of the 1990's. Initially, Kaiser Permanente placed a RFP (request from purchaser) coordinator in each of its plans. The role of the RFP coordinator was to gather and sometimes analyze data. Today, the RFP coordinators have become RFP/RFI (request for information) departments in each of Kaiser Permanente's health plans. It is estimated that Kaiser Permanente has approximately 50 individuals working on this type of activity. Each of its RFP/RFI departments has created a database to improve the efficiency of handling data requests.

In addition to the staff in Kaiser Permanente's RFP/RFI departments, the marketing and sales staff handle purchasers' requests for information. Their involvement is needed to keep up with the volume of work in this area. In order for staff from these departments to participate in these activities they must be trained to develop expertise related to performance assessment in general and HEDIS in particular. Because these staff members are also called upon to present and address performance assessment information in their roles to sell and market the plan, this type of information has been incorporated into their sales training.

Another approach Kaiser Permanente uses to deal with the challenges of performance assessment is to foster collaboration among the many professionals involved in the activities, e.g., benefit consultants, the research community and clinicians. It is critical that Kaiser Permanente staff and other professionals work together so that the data and information they provide are standardized and accurate, and that the plan's investment in this work pays off. To date, coordination of the measurement efforts in the health care industry have been less than perfect but with HEDIS, collaboration, compromise and consensus-building have increasingly become part of the process. With practice, the parties are building a joint understanding of what it takes to work together to improve the efficiency and effectiveness of their work.

In sum, national performance assessment initiatives are having a major impact on every facet of MCO beginning with their quality programs and budgets, and including staffing patterns, job descriptions, physical infrastructure, financing and payment practices and every other type of policy, procedure and practice. They are also shaping purchaser and consumer opinions and actions. What we've seen in this overview is that most national initiatives in performance assessment, to date, are process, rather than outcomes oriented. This is consistent with the formative state of the field of health outcomes. However, the infrastructures, policies, procedures and staffing developments that are taking place in MCO to accommodate current performance assessment practices are readying organizations to meet requests for outcomes information. In addition, the HOS initiative, in the early stages of implementation, represents a major step forward in patient-based outcomes in managed care. By 1999, Kaiser Permanente and other MCO will be ready to report on the impact of that measure on their organizations as well.

The next issue of the Monitor will focus on low back treatment, stroke and HIV disease. If you are working in these areas or know of an interesting project, please contact Sharon Sokoloff at (617) 426-4046 or sokoloff@outcomes-trust.org.

National Jewish Medical and Research Center, in conjunction with the Trust, is holding a conference for senior level managers in the managed care industry in Vail, Colorado from April 6-8, 1998. The conference will focus on the needs and benefits of generic and condition-specific outcomes measures in managed care. For information: Julia Jowers @ (303) 398-1083 or jowersj@njc.org.

Educational Program: Pharmacoeconomics and Outcomes
Developed by: American College of Clinical Pharmacy

Pharmacoeconomics and Outcomes is a three module, home study program that helps the participant enhance patient care by applying economic and humanistic outcomes assessment in their practice setting. It is designed for those who wish to develop and assess their knowledge in measuring and evaluating both economic and patient-based outcomes. Emphasis is placed on integrating and utilizing these methodologies in professional practice. Subscribers may earn continuing education credits. For more information visit http://www.accp.com. To order call (816) 531-2177 ext. 20. Identify yourself as a Trust member to receive a discounted price.

The Trust is pleased to announce a new member service available in February, 1998. Request a customized search of the Trust's Member database by faxing a form to Nistar DataSystems, Inc. Results will be faxed back to you within five minutes. This service will assist you with finding colleagues in the field of health outcomes assessment. We welcome your comments and suggestions. For copies of the search form contact the Trust offices at (617) 426-4046.

The Trust is developing Member Conference Call programs to provide members with small, informative, one-hour sessions to "meet the expert" or discuss certain topics with colleagues in the field. To share your ideas, please contact Sharon Sokoloff at the Trust by phone, fax or e-mail: sokoloff@outcomes-trust.org.