Image: Research
Link: Click here to skip navigational elements
Image: orange spacer
Link: Home Link: About Link: Research Link: Training Link: Dissemination/Technical Assistance
Link: What's New
Link: Health Information
Link: Library
Link: Search
Link: Tools
Link: Site Map Link: Archive
Link: Barriers and Strategies Validation Study
Link: Clinic Access Checklist Reliability Study
Link: Health Care Access Intervention Study
Link: Validation Criteria for Health Promotion Study
Link: Ratings of Health Interventions Programs
Link: Review of Health Status Measures Study
Link: Analysis of BRFSS HRQOL Study
Link: Health Status Tools Development Study
  Image: Research

R3 Health Care Access Intervention Study

Principal Investigators: David Buckley, MD; Gloria Krahn, PhD, MPH; June Isaacson Kailes, MSW, LCSW

The purpose of R3 is to test methods for improving routine health care access, especially for clinical preventive services, for people living with disabilities. Two strategies to improving access will be tested, one which focuses on patient education and empowerment and another which focuses on clinics and clinic staff. Improvements in access will be measured through clinician report, patient report, and medical encounter records of selected preventive services.

Hypothesis 1: Clinical site interventions will increase health care access (as indicated by select preventive services and health promotion activities) for people with disabilities as measured through clinician report, patient report, and medical encounter records of selected preventive services.

Hypothesis 2: Patient education and information dissemination will increase health care access for people with disabilities as measured through patient report, and medical encounter records of selected preventive services.

Hypothesis 3: Clinical site interventions combined with patient education/information dissemination will increase health care access for people with disabilities as measured through patient report, and medical encounter records of selected preventive services more than either method alone.

Study Design and Specific Aims: R3 will combine a cluster randomization trial design for the clinic intervention with patient level randomization within each clinic for the patient intervention. This design will produce four distinct categories of study participants, according to intervention exposure: 1) clinic level intervention only; 2) patient level intervention only; 3) both clinic and patient level intervention; and 4) no intervention. Sentinel measures of access to primary health care will be the preventive services of mammography, Pap test, blood cholesterol, and hemoglobin tests A1c (Hgb A1c) for diabetes (as documented in the medical claims data) as well as queries about blood pressure, weight, and smoking (as assessed by physician and patient surveys).
Specific aims are:

1) For each medical claims measure, and separately for each of the two interventions, to compare the proportion of patients who received the service before intervention to the proportion that received the service after intervention.

2) For each medical claims measure, and separately for each of the two interventions, to compare the proportion of patients who received the service in the intervention group to the proportion of patients who received the service in the non-intervention (“control”) group.

3) To survey clinicians, clinic staff members, and patients before and after the interventions to assess changes in satisfaction and in knowledge, awareness, and reported behavior regarding accessibility and utilization of routine clinical preventive and screening services for individuals with disabilities.

4) To develop regression models that describe the effect of the two interventions (individually and combined) on the performance of the preventive services measures and survey measures.

Samples: This study is conducted in partnership with the health insurance plan CareOregon. CareOregon is the largest Medicaid managed care organization in Oregon, with total enrollment of approximately 90,000 members. It does not operate its own clinics or hospitals; instead it provides coverage to enrolled members who seek care from a range of clinics and providers. The majority of participating clinics are in the Portland metropolitan area. CareOregon maintains an extensive medical encounter/claims database and health services utilization research team.

1) Study participants will be enrolled members of CareOregon, who are 18 years of age or older, of either gender, of any ethnicity, and with a Medicaid disability code for physical, cognitive, and/or sensory disabilities. About 40% of CareOregon members are of minority culture/race.

2) Clinicians and clinic staff at participating clinics constitute a participating study population. All analyses (including that of the clinic level intervention) will be at the patient level, except for the clinician and staff survey. Approximately ten primary care clinics will be enrolled to participate in the study.

In order to assure a sufficient sample size, the actual number of enrolled clinics will be determined by the number of eligible subjects in each clinic and the degree of intracluster correlation. Based on power analyses for a cluster design, we estimate that we will need a total of 85 participants in each condition to detect the anticipated effect size of 25% with .80 power and a one-tailed alpha of .05 (see Appendix B for details of power calculations used to obtain these sample sizes). This sample size will also be adequate for multivariate regression analysis, assuming a maximum of 12 independent variables. Eligibility for inclusion in the analysis of each of the four main outcomes will differ according to the nature of the particular preventive or screening service being analyzed. Individual subjects may be included in the analysis of more than one of the outcome variables, depending on their eligibility. A summary of the basic inclusion and exclusion criteria for each preventive service outcome is included in Appendix C.

Intervention, Data Collection and Measurement: The clinics will be randomly assigned in equal numbers either to receive the clinic level intervention or to serve as a “control” clinic. Within each clinic condition, eligible patients will be assigned either to receive the patient level intervention, while patients in the other half of the clinics will serve as “control” participants. Interventions are patient education through Promoting Access To Health Services (PATHS) and clinical engagement intervention (CEI).

Promoting Access to health Services (PATHS): Recruited participants will engage in a multi-step intervention process: 1) pretest data collection; 2) participation in a workshop; 3) follow-up technical assistance; and 4) post-test evaluation. The patient empowerment workshop is based on training and informational materials developed by June Isaacson Kailes. Materials will be supplemented with video demonstrations and will be standardized for research purposes. PATHS will train participants in effective methods of interacting with health care providers, including self-advocacy skills, knowledge of routine and preventive screening procedures, strategies for maximizing the quality of time spent with health care service providers, and resources for self-management of one’s condition and health. Intervention will consist of a 3-hour training session with video segments that model advocacy skills, a participant manual, and access to ongoing information and technical assistance.

Clinic Engagement Intervention (CEI): The clinic-level intervention will apply a modified community engagement process to outpatient clinical sites. The community engagement process has been used in multiple contexts (Fawcett et al., 1994) and is recognized as a “powerful vehicle for bringing about environmental and behavioral changes that will improve the health of the community and its members” (CDC/ATSDR Committe on Community Engagement, 1997). Applied to clinical sites, it represents participatory action research with health care providers, clinical administrators and patients with disabilities. Members of the RRTC: Health & Wellness staff have worked effectively with the community engagement process to increase general accessibility of communities for persons with disabilities, and have consulted on its application for improved health care access for children with special needs through Oregon’s Title V Services to Children with Special Health Needs program.
As applied to outpatient health clinics, the intervention will consist of five co-related activities: 1) meeting with clinic staff to review health care access facilitators and barriers (including review of the barriers checklist and the Americans with Disabilities Act Accessibility Guidelines (ADAAG) compliance checklist); 2) meeting with patients with disabilities to review health care access facilitators and barriers (including review of the checklists); 3) conducting joint meetings to prioritize needed changes; 4) mobilizing resources to address the needs through training, policy change, architectural or schedule changes; 5) follow-up on follow-through. Anticipated types of interventions and possible resources include architectural changes to increase physical access, provider training to increase awareness and alter behavior (e.g., Lennox, 2004), change in clinic procedures to schedule longer appointments or review charts for preventive screens for patients with disabilities, and others. This process will be facilitated by research staff who will also be responsible for providing technical assistance to the clinics in order to implement identified changes to promote accessibility.

Data Collection and Measurement: The study will collect three types of outcome data: 1) patient report of health care services and satisfaction with care; 2) health care provider report of health care services; and 3) use of preventive services measured by medical encounter data. The first measure will be a patient survey (to be developed) that asks core questions on health care, perceived accessibility and utilization of preventive services, and satisfaction with care. Data collection for the CEI will include a survey of patients with disabilities from the specified clinics. The survey will include a clinic accessibility assessment describing physical and social access issues adapted from an existing community assessment form (Appendix D) to include questions around barriers identified in study R1. The second measure will be a health care provider survey based on a modified version of our earlier questionnaire of physician practices (see Appendix E) to include additional questions on barriers identified through R1. Both the patient and provider measures may be more sensitive to early changes in practice that may not be fully reflected in medical claims records of preventive services. The third measure is medical claims data from the CareOregon database documenting use of preventive services in accordance with USPSTF guidelines. Data will consist of dichotomous variables corresponding to each of the four particular preventive services (mammography, Pap testing, serum cholesterol screening and Hemoglobin A1c (Hgb A1c) screening in diabetes), and indicating whether or not the service was performed within the recommended time period. Independent covariates to be measured for each subject include age, gender, ethnicity, language spoken, zip code, enrollment/disability status, primary care clinic, diabetes status, and hyperlipidemia status.
After both interventions have been implemented for a given clinic and its patients, a pre-intervention date (“pre-date”) and a post-intervention date (“post-date”) will be determined. These dates will be the time points at which the study data will be collected from the CareOregon database. The “post-date” will be six months after the interventions are completed, and the “pre-date” will be 12 months prior to the post-date. The study data set will be compiled by a research staff member at CareOregon, stripped of all personal identifiers, and prepared for analysis. Surveys will be administered for each clinic and its patients six months before the interventions begin and again six months after the completion of the interventions.

Data Analyses: Comparison of Groups For the survey data, multivariate analyses of variance will be used to compare intervention and control patient and provider data, as well as comparisons before and after intervention. For the claims data, analyses of proportions will be performed separately for each of the four preventive service outcomes. Using baseline data, the proportion of participants who received the preventive service within the recommended time frame will be calculated both for those participants attending intervention clinics and those attending “control” clinics. Similar calculations will be performed on “post-date” data. This will yield a measure of the proportion of those who received the appropriate preventive service for each of four groups: 1) the Clinic Intervention group; 2) the Clinic “Control” group; 3) the Patient Intervention group; and 4) the Patient “Control” group. Using the Pearson chi-square statistic adjusted for the clustering design effect, the baseline proportions will be compared to the corresponding post-intervention proportions to determine if there is a significant change in the proportion of those receiving the service after the implementation of either intervention. Finally, using only the post-intervention data (from the “post-date”), the adjusted Pearson chi-square statistic will be used to compare each intervention group with its respective “control” group.

Regression Analysis: Generalized estimating equations (GEE) will be used to fit regression models to describe the effect of the two interventions (individually and combined) on the performance of the preventive and the health promotion services. The GEE method is an approach that: can account for intracluster correlation; can accommodate most kinds of outcome variable; can account for the correlation between repeated measures on individual subjects; and can accommodate both cluster level and individual level intervention variables. This multivariate method will allow for investigation and adjustment of possible interaction effects between the principal predictor variables and/or other covariates, including the main outcome variables. For each outcome variable, plausible models will be reported, with the regression coefficient (b) and odds ratio (exp b) for each predictor variable, and corresponding p-value of the chi-square test.

Anticipated Findings: Documentation about the types of intervention strategies decided upon by local clinical sites and their effectiveness in increasing primary health care access for Medicaid enrollees with disabilities as established by three related types of data: provider report, patient report and medical claims data. Findings will also document the effectiveness of training adults in patient empowerment education, and its relative additional contribution to clinic site intervention. These findings will include effective strategies to improving health care access and use of preventive services.





We want your feedback! Please tell us what you think of the information on this page.

Take the short survey!

Home | About | Research | Training | Dissemination/Technical Assistance | Health Info | Library
Search | Tools | Site Map | What's New

© 2004, RRTC Health & Wellness Consortium. All Rights Reserved. Contact webmaster.

Updated 4/15/05