Disseminating quality improvement: Study protocol for a large cluster-randomized trial

Andrew R. Quanbeck, David H. Gustafson, James H. Ford, Alice Pulvermacher, Michael T. French, K. J. McConnell, Dennis McCarty

Research output: Contribution to journalArticlepeer-review

23 Scopus citations


Background: Dissemination is a critical facet of implementing quality improvement in organizations. As a field, addiction treatment has produced effective interventions but disseminated them slowly and reached only a fraction of people needing treatment. This study investigates four methods of disseminating quality improvement (QI) to addiction treatment programs in the U.S. It is, to our knowledge, the largest study of organizational change ever conducted in healthcare. The trial seeks to determine the most cost-effective method of disseminating quality improvement in addiction treatment.Methods: The study is evaluating the costs and effectiveness of different QI approaches by randomizing 201 addiction-treatment programs to four interventions. Each intervention used a web-based learning kit plus monthly phone calls, coaching, face-to-face meetings, or the combination of all three. Effectiveness is defined as reducing waiting time (days between first contact and treatment), increasing program admissions, and increasing continuation in treatment. Opportunity costs will be estimated for the resources associated with providing the services.Outcomes: The study has three primary outcomes: waiting time, annual program admissions, and continuation in treatment. Secondary outcomes include: voluntary employee turnover, treatment completion, and operating margin. We are also seeking to understand the role of mediators, moderators, and other factors related to an organization's success in making changes.Analysis: We are fitting a mixed-effect regression model to each program's average monthly waiting time and continuation rates (based on aggregated client records), including terms to isolate state and intervention effects. Admissions to treatment are aggregated to a yearly level to compensate for seasonality. We will order the interventions by cost to compare them pair-wise to the lowest cost intervention (monthly phone calls). All randomized sites with outcome data will be included in the analysis, following the intent-to-treat principle. Organizational covariates in the analysis include program size, management score, and state.Discussion: The study offers seven recommendations for conducting a large-scale cluster-randomized trial: provide valuable services, have aims that are clear and important, seek powerful allies, understand the recruiting challenge, cultivate commitment, address turnover, and encourage rigor and flexibility.Trial Registration: ClinicalTrials. govNCT00934141.

Original languageEnglish (US)
Article number44
JournalImplementation Science
Issue number1
StatePublished - Apr 27 2011

ASJC Scopus subject areas

  • Health Policy
  • Health Informatics
  • Public Health, Environmental and Occupational Health


Dive into the research topics of 'Disseminating quality improvement: Study protocol for a large cluster-randomized trial'. Together they form a unique fingerprint.

Cite this