Value in Development of Complex Interventions

May 1, 2006
Anne E. Sales, PhD, RN

Christian D. Helfrich, MPH, PhD

Volume 12, Issue 5


The article by Byrne et al1 in this issue of theoffers a more systematic approach toplanning complex interventions than has generallybeen used in the past. It is an important addition tothe literature on planned and organized diffusion ofinnovation and implementation of evidence-based care.The authors provide an excellent counterpoint to prescriptiveapproaches for designing interventions tospeed the adoption of evidence-based practice.Implementation studies too often focus on specific tools,such as reminders, provider education, and performancefeedback; treating contextual factors, such asorganizational structure and culture; and the healthpolicy environment, as factors to be controlled throughrandom assignment.2-4 Such studies tell us neither whyinterventions achieve the results they do, nor why wemight expect different results from one setting toanother. If there is consensus in implementationresearch, it is that there is tremendous unexplainedvariation in effectiveness of diffusion, adoption, andimplementation.5-10

Understanding these differences is a goal of manyhealthcare organizations and funders of health servicesresearch. Indeed, the Institute for Healthcare Improvement'sacclaimed "Break-through Collaboratives" adoptedthe plan-do-study-act approach in part to addressdifferences in context: to try out multifaceted changeinterventions on a small scale in order to understand howthey work in a given context and to provide an opportunityfor refinement before rolling them out on a largerscale. The Medical Research Council (MRC) frameworkserves a similar purpose, but provides more structure,particularly in beginning with a systematic review of theexisting literature, focus groups, and pilot testing.

In secondary prevention of coronary disease, theauthors find an excellent example for applying andstudying the MRC framework. These therapies are relativelysimple to deliver. While not appropriate for allpatients, prescribing the "cardiac cocktail" is not ascomplex as applying evidence-based therapies in otherdisease states. In addition, there have been multipleefforts to implement adherence to evidence-based therapiesfor coronary disease prevention across settings, inmany healthcare systems, and in several countries. Yet,despite the relative ease of adherence (compared totherapies in other healthcare problems) and the amountof attention that has been focused on the area, we havenot yet found consistent, transferable approaches toincreasing adherence to these practices across settings,systems, and countries. This makes it an excellent clinicalarea in which to launch complex interventions: Theevidence base is strong, and the therapies are relativelysimple to administer.

Why This Study Is Relevant to US Managed Care

US readers may look at this study's Irish setting andpresume the study has little bearing on their own work.This would be incorrect for 2 reasons.



First and foremost, the MRC framework is not aboutIrish healthcare system so much as it is aboutaccounting for the dynamics of given healthcare settingwhenever designing a trial. As a result, the frameworkis highly applicable to the American healthcaresystem, which has enormous variation in environmentsfrom locality to locality. Differences arise from state regulations,levels of managed care penetration, and extentof employer-sponsored coverage, not to mention differencesin demographic factors such as age and ethnicity,and behavioral factors such as regional differences indiet and lifestyle. All of these factors profoundly affecthow a complex intervention might work in a given setting.A major purpose of the MRC framework is to capturecontextual variables like these in a systematic way.

Second, the findings of this study may have more relevanceto the US healthcare system than one mightassume initially because the healthcare system inIreland is quite similar to that in the United States insome important regards. First, general practice physiciansare not employees of a public health service, butindependent practitioners who provide services undercontract to public agencies called Health ServiceExecutive Areas. As a result, the discussion related tophysician willingness to participate and barriers to complyingwith evidence-based practice is likely to be similarto responses and attitudes of US primary carephysicians. Second, Ireland does not have universalhealth insurance coverage for all services, particularlyoutpatient primary care; the majority of Irish citizensare covered for these services under private arrangementswith either employment-based insurance coverageor out-of-pocket payment.

There are differences. In Ireland lower income residents,older persons, and other groups are guaranteedcare and all citizens are guaranteed certain services,such as public healthcare and maternity care. This hasresulted in the creation of public bodies that plan andoversee healthcare services in ways quite unlike in theUnited States. As a result, the perception that providingpreventive services, including secondary prevention, asa public good may be stronger in Ireland than in theUnited States, and public sector bodies, in particular,may have a greater willingness to fund interventions toimprove preventive care. While different from much ofthe US healthcare system, this support for preventivecare is very similar to the few public systems of carein the United States, notably the Veterans HealthAdministration of the Department of Veterans Affairs. Itis also relevant to health management organizations(HMOs) that include both delivery and payment systemsunder the same organization, such as staff modelHMOs. In these organizations, similar economic andpolicy incentives exist that align healthcare and publichealth interest in longer term outcomes.

Overall Lessons

The complex intervention development described inthe article by Byrne et al functions as an important"ideal case" of following a highly structured, intensiveprocess of preparing interventions for systematic study.Implicit in the MRC framework is an investment of timeand resources at the initiation of an implementationeffort, in order to understand the contexts in which theintervention will be implemented and to maximize theprobability of success. This is relevant to the market-basedUS healthcare system, as researchers and healthcareproviders in the United States have myriadcontextual factors unique to the local environment withwhich to contend. The MRC framework provides astructured method of assessing those factors. To be ofgreatest utility, this article should be read in conjunctionwith the articles describing the outcomes of therandomized trial for which Byrne et al applied the MRCframework.

From the VA Puget Sound Health Care System, Seattle, Wash.

Address correspondence to: Anne E. Sales, PhD, RN, Health Service Research andDevelopment, VA Puget Sound Health Care System, 1880 S. Columbia Way, Seattle, WA98108. E-mail:

Am J Manag Care.

1. Byrne M, Cupples ME, Smith SM, et al. Development of a complex interventionfor secondary prevention of coronary heart disease in primary care using the UKMedical Research Council framework. 2006;12:261-266.

Jt Comm J Qual


2. Solberg LI, Brekke ML, Fazio CJ, et al. Lessons from experienced guidelineimplementers: attend to many factors and use multiple strategies. 2000;26:171-188.

Eff Clin Pract.

3. Solberg LI, Kottke TE, Brekke ML, et al. Failure of a continuous quality improvementintervention to increase the delivery of preventive services. A randomizedtrial. 2000;3:105-115.

Eff Clin Pract.

4. Solberg LI, Kottke TE, Brekke ML, Magnan S. Improving prevention is difficult.2000;3:153-155.



5. Greenhalgh T, Robert G, Macfarlane F, Bate P, Kyriakidou O. Diffusion of innovationsin service organizations: systematic review and recommendations. 2004;82:581-629.

Health Aff (Millwood).

6. Shojania KG, Grimshaw JM. Evidence-based quality improvement: the state ofthe science. 2005;24:138-150.

Am J Med.

7. Shojania KG, Grimshaw JM. Still no magic bullets: pursuing more rigorousresearch in quality improvement. 2004;116:778-780.

Med J Aust.

8. Grimshaw JM, Eccles MP. Is evidence-based implementation of evidence-basedcare possible? 2004;180(suppl 6):S50-S51.

Med J Aust.

9. Sanson-Fisher RW, Grimshaw JM, Eccles MP. The science of changing providers'behaviour: the missing link in evidence-based practice. 2004;180:205-206.

Health Technol Assess.

10. Grimshaw JM, Thomas RE, MacLennan G, et al. Effectiveness and efficiency ofguideline dissemination and implementation strategies. 2004;8:iii-iv, 1-72.