Table 2.

Description of examples for implementing and spreading interventions into multilevel contexts^*

Example	Intervention levels	Setting(s)	Design	Data collection	Measures
Pool Cool Diffusion Trial (skin cancer prevention program) (39,49)	Field coordinators Swimming pools (recreation centers) Children (ages 5–10 y)	32 metropolitan regions across United States, each responsible for cluster of 4–15 pools per coordinator	Three-level nested experimental design with field coordinators randomized to Basic and Enhanced (reinforcement plus feedback) diffusion conditions over 3 years Process evaluations among key stakeholders at multiple levels	Pool managers surveyed each summer Sample of parents at each pool surveyed about their children (baseline and end of each summer) Cohort subsample followed over multiple years Interviews and formative progress reports	Pool-level implementation, maintenance and sustainability over successive years Organizational/environmental change at pools Child sun-protection habits and sunburns
CHOICE (40,50)	Practice-level (physician practices with Aetna HMO and ≥50 members ages 52–75 y) Patient-level (health-plan members)	32 primary care practices in FL† and GA† and 443 participating health plan members (211 in intervention practices, 232 in usual care practices)	Cluster randomized controlled trial Modified blocked randomization, stratified on practice size, number of eligible members and urban/rural Augmented enrolled patients (eligible per completed baseline survey) with nonresponders (for whom claims data were reviewed)	Practice surveys (before and at end of study) Patient surveys (eligibility, baseline and follow-up surveys at 12 and 18 mo) (multimodal, ie, mail, telephone, web) Claims data (CRC screening in addition to patient surveys)	Physician practices’ CRC screening practices, referrals and quality improvement initiatives Patient CRC screening test completion Patient interest in CRC screening, intent to ask providers about screening, readiness to be screened
Improving Systems for Colorectal Cancer Screening (51,52)	Integrated medical group (HVMA) Primary care physicians Adults (ages 50–80 y) who were overdue for screening	Multispecialty, integrated medical group serving approximately 300 000 patients at 14 clinical centers in Eastern MA†	2 × 2 design Approximately 22 000 patients randomized to receive mailed reminders and FOBT cards 110 primary care physicians randomized to receive patient-specific electronic clinical reminders during office visits for patients who were overdue for CRC screening Subsequent randomized study of electronic reminders and web-based CRC risk index from primary care physicians to approximately 1100 patients who were overdue for CRC screening and were registered users of web-based patient portal in EMR	EMR data Physicians surveyed after the intervention	Patient completion of FOBT testing, flexible sigmoidoscopy or colonoscopy over 15 mo Detection of CRC or adenomas among adult patients Physician views of CRC screening and value of electronic reminders
Best Practices for Comprehensive Tobacco Control Programs (45,53)	National policy (government and advocacy groups) State policy State tobacco control programs Local initiatives Community-based settings	State and local departments of health and education Schools Worksites Restaurants Bars	Multiple studies with different designs at different levels (eg, time series and multiple baseline studies) Individual state-level evaluations National evaluation Federally-funded intervention studies (eg, COMMIT, ASSIST, IMPACT, NTCP) Cost–benefit and cost-effectiveness analyses (eg, if children protected from secondhand smoke, could prevent more than a million asthma attacks and lung and ear infections)	Local, state, and national surveillance (eg, BRFSS, national and state Adult Tobacco Surveys, Youth Tobacco Surveys, Current Population Survey Tobacco Use Supplement) Individual intervention studies on tobacco use screening, brief interventions by clinicians, quitlines, tobacco use treatment and coverage, etc. Media studies tracking appearance of tobacco news and other items in popular press, counter-advertising coverage	Tobacco consumption or cessation ratesCigarette sales Heart disease deaths Lung cancer incidence Hospitalization ratesTobacco control program spending Coverage of tobacco issues/developments in popular media (content analyses of news stories) Use of quitlines and other cessation services Policy changes Changes in social norms Legislative tracking
TIDES (46–48)	Primary care practices and their providers Associated local medical care system leadership and mental health specialists Regional system interdisciplinary leadership National research–based QI organization (QUERI) National system patient care services policy makers	VA medical centers and community-based outpatient clinics Early pilot in 1 VA practice Trial in 10 practices Spread to >60 practices National implementation	Multiple studies with different designs, all with purpose of providing formative information to lead to effective national policy All based on collaborative care model tested in multiple randomized trials and validated in meta-analysis Cluster randomized trialQualitative data and analysis at the levels of the care model evidence-based QI intervention,and the national policy intervention Prepost testing of individual patients participating in the collaborative care intervention	Consecutive primary care patients screened for depression and tested for process of care and clinical outcomes Local primary care practices, including nurses, primary care clinicians, and mental health specialists directly observed and interviewed Regional and national policy activities tracked Primary care providers surveyed Costs of the intervention and implementation tracked	Qualitative themes Rates of initiation of antidepressants Depression symptoms Functional status Utilization (primary care and mental health visits, hospitalizations) Provider knowledge, attitudes, behaviors (including comfort with managing depression, degree of collaboration) Practice-level measures of structures, processes and implementation
VHA Colorectal Cancer Care Collaborative (C4) Cancer Care Quality Measurement System (43,54)	National policy National quality and performance managers Regional leadership (VA networks) Medical centers (leadership and quality managers) Community-based outpatient clinics Primary care practices and their providers	VA medical centers and community-based outpatient clinics Initial studies single site chart review and national data analyses Collaborative comprises 21 medical centers (one from each VA network) Best practices disseminated to all medical centers	Quarterly audit and feedback of chart-based performance measures Comparative facility-level analyses within and across VA networks	Externally conducted chart reviews of CRC screening among random samples of eligible veterans Electronic data abstraction tool into which teams enter patient-specific data for CRC treatment measures (about 230 data elements used to evaluate compliance with 24 quality indicators)	CRC screening: CRC screening for all average or high-risk patients60-d follow-up diagnostic tests of positive screens CRC treatment: Treatment indicators (eg, proportion of patients with T3 or T4 lesions and/or node+ disease referred to medical oncologists); timeliness of treatment processes (eg, dosing/timing of chemotherapy, adjuvant radiation); postsurgical surveillance

^*ASSIST = American Stop Smoking Intervention Study for Cancer Prevention; BRFSS = Behavioral Risk Factor Surveillance System; CHOICE = Communicating Health Options through Information and Cancer Education; COMMIT = Community Intervention Trial for Smoking Cessation; CRC = colorectal cancer; EMR = electronic medical record; FOBT = fecal occult blood test; HMO = health maintenance organization; HVMA = Harvard Vanguard Medical Associates; IMPACT = Initiatives to Mobilize for the Prevention and Control of Tobacco Use Program; NTCP = National Tobacco Control Program; QI = quality improvement; QUERI = Quality Enhancement Research Initiative; TIDES = Translating Initiatives in Depression into Effective Care Solutions; VA = US Department of Veterans Affairs health-care system; VHA = Veterans’ Health Administration.

^†32 primary care practices included in CHOICE study are in Florida (FL) and Georgia (GA). 14 clinical centers included in Improving Systems for Colorectal Cancer Screening study are in Massachusetts (MA).