Maternal and Child Health Journal, Vol. 2, No. 4, 1998

Adopting Immunization Recommendations: A NewDissemination Model

Gary L. Freed, M.D., M.P.H.,1,4 Donald E. Pathman, M.D., M.P.H.,2,3Thomas R. Konrad,Ph.D.,3 Victoria A. Freeman, R.N., Dr. P.M.,3 and Sarah J. Clark, M.P.H.1

Objective: This paper presents a new approach for understanding factors related to physicianadoption of clinical guidelines, using children's vaccine recommendations as a case study.Methods: The model traces sequential steps, from awareness to agreement to adoption and,finally, adherence to the guideline. Movement through these stages can be catalyzed orretarded by many influences, grouped into two major categories: environmental characteris-tics of the physician's practice, and information characteristics of the guideline. Environmen-tal characteristics include sociocultural factors, professional characteristics, and practice orga-nization factors. Information characteristics include the guideline's relative advantage,complexity, and compatibility with existing guidelines and protocols, as well as mechanismsof guideline dissemination. Implications: This model can be used to identify characteristicsthat will likely impede or facilitate guideline adoption, and to focus dissemination effortson key issues.KEY WORDS: immunization; guidelines; dissemination; recommendations.

INTRODUCTION

There are many reasons that U.S. children donot receive their immunizations at the appropriateages, including difficulties with access to care. How-ever, even when children successfully gain access,health care providers frequently fail to offer and pro-vide immunizations (1, 2). To deal specifically withthis problem, the Healthy People 2000 report set asa goal to "increase to at least 90% the proportion ofprimary care providers who. . . offer immunizationsas appropriate to their patients" (3).

The most publicized issues related to provider1Division of General Pediatrics, University of Michigan HealthCare System, Ann Arbor, Michigan.

2Department of Family Medicine, University of North Carolinaat Chapel Hill, Chapel Hill, North Carolina.

3Cecil G. Sheps Center for Health Services Research, Universityof North Carolina at Chapel Hill, Chapel Hill, North Carolina.

4Correspondence should be directed to Gary L. Freed, M.D.,M.P.H, Division of General Pediatrics, University of MichiganMedical Center, 1500 E. Medical Center Dr., D3202 MPB, AnnArbor, Michigan 48109-0718; e-mail: gfreed@umich.edu

immunization practice involve inappropriate contra-indications and missed opportunities for immuniza-tion. Often overlooked is the question of whetherphysicians who care for children understand and fol-low immunization recommendations.

In ail areas of medicine, the number of clinicalpractice guidelines put forth has increased substan-tially in recent years. Yet relatively little attentionhas been paid to the issue of how to disseminateclinical practice guidelines to influence patient caremost effectively. Unfortunately, the work publishedto date has had little impact improving disseminationmethods (4). As a result, many guidelines remainunfamiliar to physicians and have little effect on clini-cal practice (5, 6). In this paper, we apply the caseof children's vaccine recommendations to a new para-digm for understanding the factors related to physi-cian adoption of clinical guidelines. We believe exam-ination of this case not only illustrates some of therecurrent concerns that are relevant to the formula-tion and implementation of clinical policies in a vari-ety of settings and clinical domains, but also serves

2311092-7875/98/1200-231$15.00/0 C 1998 Plenum Publishing Corporation

232 Freed et al.

as an ideal vehicle to enrich our theoretical under-standing of the processes underlying guideline dif-fusion.

UNDERSTANDING GUIDELINEDISSEMINATION

There are five reasons why immunization recom-mendations provide such an important test case forunderstanding guideline dissemination:

1. Until fairly recently, there had been a longperiod during which a uniform, universal im-munization schedule was in place and thoughtto be generally accepted by most physicians.Since physician practice rarely changes in thisarena in the absence of a new guideline,changes can be assessed from a period of rela-tive stability and temporally associated withthe introduction or modification of specificguidelines.

2. Considered in the context of other preventiveservices delivered by physicians, immuniza-tion is a "medicalized" procedure unlike, forexample, dietary counseling and behavioralcounseling, which have a more significant psy-chosocial orientation. Physicians are consis-tently and broadly trained to give vaccines—amaneuver that is relatively simple and clearlyconsidered a typical part of medical practice.

3. The administration of children's vaccines is arelatively discrete, unambiguous activity thatgenerally is documented in the ambulatorycare medical record, unlike many preventiveservices. One of the sequelae of the concernwith liability over vaccines is that the adminis-tration (or non administration) of a vaccineleaves a paper trail.

4. Generally, immunization recommendationsare universal or, occasionally, limited to spe-cific geographic areas where epidemics haveoccurred. Because contraindications for vac-cine administration are few and usually areexperienced temporarily, it is relatively easyto determine which immunizations should begiven at a certain age compared to more com-plex contraindications for many other proce-dures.

5. Finally, immunization delivery is being incor-porated into regimens designed to assess ofthe quality of care delivered by physicians,

and to rate the effectiveness of health careplans. Yet the administrative complexitiesand practical challenges involved in the mea-surement of this apparently straightforwardindicator poses important challenges to ef-forts aimed at assessing more complex indi-cators.

Despite the importance of studying the dis-semination of immunization recommendations tophysicians, little is known about this process or thedeterminants of recommendation adoption and ad-herence. Traditionally, physician agreement andadoption of new immunization recommendations wastaken for granted by medical societies and publichealth officials. However, recent studies demonstra-ting variation in awareness and adoption of new im-munization recommendations for Hib and hepatitisB vaccines emphasize that adoption can no longerbe assumed to occur uniformly (7-11).

Each new or revised immunization recommen-dation increases the complexity of immunization de-livery and requires a change in practice behavior.Moreover, the process of disseminating and cultivat-ing compliance with new immunization recommen-dations has been compounded by the existence ofseveral confusing, competing, and sometimes contra-dictory vaccine administration schedules issued bydifferent professional and public health bodies. Tocorrect these inconsistencies, uniform immunizationschedules have been released as a joint effort of theAdvisory Committee on Immunization Practices(ACIP), the Centers for Disease Control and Preven-tion, the American Academy of Pediatrics, and theAmerican Academy of Family Physicians (12). How-ever, the initial approval of new recommendationsstill is not coordinated, as evidenced by recent experi-ence with varicella vaccine and injectable polio vac-cine. There continue to be periods of uncertainty forproviders in which ACIP recommendations do notmatch those of their own professional association.

A CONCEPTUAL MODEL OF GUIDELINEADOPTION AND ADHERENCE

To provide a framework for research into thedeterminants of immunization guideline adherencefor the individual practitioner, we constructed andtested (13) a conceptual model of this process(Fig. 1). Since that time, we have developed a greaterunderstanding of the complexity, variety, and interre-

Adopting Immunization Recommendations 233

Fig. 1. Model of the immunization recommendation implementation process.

lationships of the variables affecting each stage ofthe process. We have performed additional studiesexploring further the impact of parental factors (14),policy changes (15, 16), and legal issues (17, 18) onthe adoption and adherence to immunization recom-mendations. The model presented here is now morecomplete in its demonstration of the process and itsdeterminants with sequential stages from preaware-ness through implementation and adherence. Ac-cording to the theoretical base of the model, move-ment through each stage in the process can becatalyzed or retarded by such factors as physiciancharacteristics (19-21), practice organization (22,23), practice environment (24), and characteristics ofthe vaccines and the guidelines themselves (25). Thismanuscript further details these factors and identifiesthe stage(s) at which they have an effect on theprocess.

The many influences on this process are orga-nized in the model to better understand their com-plexity and interrelationships. Two broad categoriesencompass all others: environmental characteristicsof a physician's practice setting and information char-acteristics of the recommendation.

Environmental Characteristics

Environmental characteristics include sociocul-tural, professional, and organizational factors. Eachof these can be broken down further to specific fac-tors that have an impact on each stage in the process.

Sociocultural factors, seen in the top portion ofFig. 1, include such things as previous history of epi-demics, community or school vaccination policies,and parent demand. Professional and organizationalfactors include individual physician characteristics(e.g., age, specialty), practice characteristics (e.g.,solo, group), and practice management policies.

Individual Physician Factors

Considerable attention has been given to therole played by features of individual physicians inmedical innovation diffusion. Eisenberg concludedthat the most important factors influencing clinicaldecision making were personal characteristics of indi-vidual physicians and their interactions within theirprofession (19). Individual physician factors, includ-

234 Freed et al.

ing age and training, have been carefully studied inregard to the use of laboratory tests and prescriptionof drugs (20, 21, 26). Although individual physiciancharacteristics have been found to be associated withpractice patterns and prescribing styles, these associa-tions vary with the class of drugs being studied. Linnfound that prescribing patterns of psychotropic drugswere more likely to be related to characteristics re-flecting physicians' "values, social position, or socialbackground" than indicators of medical or scientificbackground (27, 28). Stolley noted that physiciansmore appropriate in their chloramphenicol prescrib-ing habits were those who were younger, "more cos-mopolitan," more modern, and those concerned withboth the psychosocial and quality dimensions of med-ical care (21). The specific factors influencing medicaldecision making have been found to vary among spe-cialties (29-31), perhaps related to known specialtydifferences in individuals' styles of information shar-ing and patterns of adoption of new treatments (32).This variation across specialties only complicates ourunderstanding of the influence of individual factorson medical practices.

Burt categorized physicians into four groupsbased upon the timing of their adoption of tetra-cyline: eager innovators, early conformers, lateconformers, and deviant laggards (33). Althoughassociations were noted between some physiciancharacteristics (e.g., age, number of journal subscrip-tions maintained, early contact with pharmaceuticalrepresentative) and the date of tetracyline adoption,Burt reported no attempt to specifically identify indi-vidual characteristics in each of these groups.

Practice Specific Characteristics

It is increasingly recognized that medicine is notpracticed in a vacuum, but exists within a complexnetwork of intra- and interorganizational arrange-ments influencing physician practice patterns (22).Thus, the ability to influence physician decision mak-ing must be based on our ability (a) to understandand affect the organizational context within whichthe physician functions, and (b) to recognize the net-work of organizations that influence activities withinthe practice (34).

Becker's data suggests that opinion leaderswithin health care organizations are selective in theirsupport of innovations (35). Mohr speculated thatindividual motivation to innovate will be successfulin direct relation to the availability of supportive or-

ganizational resources in a given situation, and ininverse relation to organizational obstacles to innova-tion (36). These factors may overpower individualefforts to adopt or reject new medical treatments. Theeffects of group style and peer pressure are thought tobe stronger in more formally organized practices,such as health maintenance organizations (37). Wil-liamson noted that the average time from awarenessto adoption of new therapies was shorter for physi-cians in group practices than those in solo practice(23). Specifically, increasing time spent with col-leagues was related to earlier adoption. The interac-tion between physicians, as fostered in medicalgroups, appears to enhance dissemination of informa-tion about new therapies, and helps legitimize aninnovation (23,38). Eisenberg noted the strong effectof "educationally influential" physicians within prac-tices (37).

Other aspects of organizations that play signifi-cant roles in the diffusion of medical information andtechnology adoption include organizations' formalquality assurance efforts, methods used to enforcegroup standards of care, information managementsystems, patient flow efficiency, and delegation ofpatient care tasks to nurses and aides. These concretefactors, which reflect Mohr's supportive organiza-tional resources or obstacles, (36) have received littleformal assessment.

Organizational policies and procedures probablyplay a strong role in the ability of physicians withina practice to achieve high immunization rates amongtheir pediatric patients. Appropriate times for well-child visits to physicians (tied to the ages when vac-cines are to be given) are widely accepted, and likelyassist in achieving high vaccine rates. It is a nearuniform practice to record in a specially designedlocation in pediatric patient charts all vaccines admin-istered, along with manufacturer batch information,and dates of administration. The availability of thisdata has helped make vaccine compliance a favoritetarget for practice's quality assurance efforts.

Practice Management Policies

Physicians, though operating within organiza-tions, also interact with and are influenced by outsidephysicians. Coleman, in a study of tetracycline adop-tion soon after its release, determined that the behav-ior of a physician's peers was the strongest determi-nant of practice behavior (39). Those who were lateadopters had peers who were late adopters. The in-

Adopting Immunization Recommendations 235

fluence of physicians outside one's practice is likelygreatest for physicians in smaller groups or solopractice.

Physicians' medical practices are influenced in-creasingly by outside forces. Medical care decisionsare guided by the requirements of health care payorsfor second opinions for surgical procedures, and limi-tations in drug prescribing options through restrictiveformularies. Carey and Weis suggest that despite lowvaccination rates in general for medicaid patients,that those enrolled in capitated arrangements mightfare somewhat better (24).

State or local regulations specifying vaccine re-quirements for public school attendance are felt toexert strong influences on physicians' vaccine prac-tices (17). Some state and local health departmentsare called upon to set local vaccine recommendationsto meet recent vaccine-preventable disease out-breaks. They then also serve to educate physiciansabout these recommendations.

Information Characteristics

Information characteristics comprise specificfeatures of the guideline itself and the mechanismsof the dissemination process. Information character-istics unique to each immunization recommendation,seen in the lower portion of Fig. 1, might involve therelative advantage of the revised or new recommen-dation, its complexity for providers and parents, andits compatibility with existing recommendations. Themechanisms of dissemination involve the choice ofinformation sources (e.g., medical or lay, official orunofficial), the credibility of those sources, othermethods of dissemination, and the initial acceptanceby local medical opinion leaders.

Guideline Specific Factors

Innovation theory has taught that certain fea-tures of innovations themselves enhance their likeli-hood of being adopted (25,30). Three specific charac-teristics—relative advantage, complexity, andcompatibility—are found to be the most reliable pre-dictors of adherence (40). Relative advantage refersto the degree to which the innovation is perceivedas being better than that which it is intended to re-place. Two specific dimensions of relative advantageare the innovation's scientific performance and cost.Complexity refers to the degree to which the innova-

tion is difficult to understand or use. Compatibilityis the degree to which the innovation is seen as con-gruent with existing modes of operation, values,and needs.

Vaccine practices are influenced by characteris-tics of specific recommendations such as the timingrecommended or the way in which contraindicationsare identified, or even the wording used in theserecommendations. The notions of relative advantage,complexity, and compatibility seem highly relevantto explaining variation in vaccine protocol adherence.The past success of vaccine recommendation compli-ance may be explained by vaccines' perceived relativeadvantage (clear efficacy in disease prevention), lim-ited complexity (recommendations traditionally easyto understand and injections themselves easy to ad-minister), and high compatibility with medical officepractice (injections are well accepted by parents pa-tients and fall well within the traditional scope ofwork of physicians). An example of a successful in-stance of implementation of an innovation is the caseof the change in Haemophillus influenza, type b, rec-ommendations where adoption by physicians was rel-atively rapid (10). The relative advantage offered bythe new recommendation was quite easy to conveyto physicians, given the promise of disease protectionthrough the earlier months of life carrying the great-est morbidity and mortality risk. Further, compatibil-ity was high, given that the added vaccine doses wereto be given at the same 2-, 4-, and 6-month well childvisits as recommended for diptheria, pertussis, andtetanus (DPT). Finally, complexity was low for thisrecommendation change since the major issue wasmerely a change in the timing of the first vaccine.

On the other hand, growing resistance of physi-cians and patients to recent and upcoming vaccinerecommendations can be understood as emerging ina milieu that is itself changing. With the emergenceof new vaccine recommendations for diseases per-ceived as less severe or life-threatening (e.g., rotavi-rus, varicella) (14), relative advantage will be lessclear, complexity will increase, and compatibility can-not be taken for granted. The perception of relativeadvantage of new recommendations wanes as theliving historical memory of dreaded childhood dis-eases diminishes, affecting the judgments of parentsand physicians. Further, additional vaccines involveadditional costs to parents and payors for children'shealth maintenance. On the other hand, physiciansdid not initially manifest consensus about the relativeadvantage of universal vaccination for hepatitis B(8, 9, 41), given that it is a disease with well-known

236 Freed et al.

lifestyle risk factors. Each new vaccine recommenda-tion or modification increases the complexity of theentire process. As more and more vaccines are addedto the schedule, room for confusion and error grows.

Guideline Dissemination Factors

Relatively little research has been devoted tothe question of dissemination of guidelines in the areaof vaccines. In a study of 12 organizations (HMO's,insurers, specialty societies) that develop and usepractice guidelines, Audet, Greenfield, and Field sur-veyed a variety of professional, commercial and prac-tice based organizations who were developing of dis-seminating medical practice guidelines (42). Theyfound that less emphasis was placed on implementa-tion than on guideline development. In general, orga-nizations that both develop and use guidelines (i.e.,payers and delivery organizations) appeared to de-vote more attention to implementation than did pro-fessional organizations (i.e., medical specialty socie-ties). Less systematic dissemination of guidelinesmade updating of guidelines more difficult. Addingto the complexity is the discovery that different spe-cialties have varying preferences for the source ofnew vaccine recommendations (43, 44).

In total, the process appears complex and convo-luted, suggesting the difficulties encountered in dis-seminating new guidelines to achieve widespread ad-herence. The following sections examine each stepin the process and the particular influences that mayhave an impact upon them.

Awareness. Before any publicity is releasedabout a new recommendation, physicians are in astate of preawareness. The process of implementinga new immunization recommendation begins withawareness, where physicians learn about the guide-line. The sources of this new information may includecommunications from the Centers for Disease Con-trol, specialty societies, and local health departments,articles in medical journals, or television and otherlay media efforts. Colleagues are also a source ofinformation for many practitioners. The mechanismof dissemination unique to these sources may havea direct impact on their success in promotingawareness.

The individual characteristics of physicians alsoplay a role in awareness. Physicians of different ages,specialties, and professional society affiliation mayhave differential access to information about a newrecommendation. They vary in the number of meet-

ings attended, the amount of time spent reading med-ical literature, and the attention devoted to a particu-lar medical issue (in this example, childhoodimmunization).

Agreement. The next step in the process,agreement, occurs when a physician conceptually un-derstands and approves of the rationale for a recom-mendation. Movement toward agreement is influ-enced by a combination of factors, most importantly(1) credibility of the information source, (2) the rela-tive advantage of the new recommendation over thestatus quo, (3) the complexity of the new recommen-dation, and (4) agreement by local medical leaders.Credibility of the information source is essential infostering trust in the information presented. Con-cerns about liability and patient care will make physi-cians hesitant to accept new recommendations fromsources not known for their accuracy or complete-ness. Increasingly, sources of information about im-munization recommendations must "prove theircase"—i.e., offer convincing evidence that the newor revised recommendation offers an advantage overthe existing schedule. The importance of this factorwas demonstrated by low rates of physicianagreement with universal infant hepatitis B immuni-zation (8, 9).

Previous research also has demonstrated the rolethat local medical opinion leaders play in fosteringwidespread agreement with innovation (37). Fre-quently, there are respected physicians in the commu-nity who guide colleagues in establishing communitynorms and determining standards of care. Their pub-lic agreement with a new recommendation may pro-mote agreement by others (35).

Physician and practice characteristics also influ-ence agreement. For example, members of specialtysocieties may be more likely to agree with a recom-mendation from their own organization. Also, ratesof agreement with new recommendations have beenshown to differ by practice type (e.g., solo, group,public, HMO) (8). This may reflect the nature ofthe individual physician who gravitates to a certainpractice type.

Adoption. The next step is adoption, where phy-sicians move from an abstract and intellectual con-cept (agreement) to a concrete endorsement of therecommendation to patients and preparations to alterclinical practice accordingly. Environmental factorsinvolved in achieving this stage include the size andcomplexity of the practice organization, physiciancharacteristics, and local or state vaccine require-ments for school entry. Informational factors include

Adopting Immunization Recommendations 237

the complexity of the recommendation, its relativeadvantage over current recommendations, and itscompatibility with existing recommendations (45).

Elements of a physician's practice setting ofteninfluence adoption (46). Physicians in small, indepen-dent, solo or two-physician practices may have littledifficulty adopting new immunization recommenda-tions; the process may be as simple as ordering a newproduct and training a small number of office staff.In a larger practice or HMO, the process of adoptionoften requires administrative approval from a clinicalpractice committee or from a third-party payer. Inthis setting, if approval is given, adoption of the newpolicy would be rapid and widespread, affecting manyphysicians. Conversely, if approval is denied, few ifany physicians in the organization would adopt thechange. Additionally, outside influences, such as re-quirements for school or day care attendance set bylocal or state ordinance, can foster and even hastenadoption of a new recommendation by influencingpatient demand even when agreement with it isminimal.

The more complex a new recommendation, thelonger it likely will take to adopt. Physicians mustdetermine the best method to put a recommendationinto action, often based on logistical concerns, (e.g.,record-keeping, scheduling). The compatibility of thenew recommendation with the immunization sched-ule already in place can also impact on the rapidityof adoption. For example, a new recommendationthat necessitates additional office visits imposes manymore difficulties than one that substitutes or addsinjections at the regularly scheduled well-child visits.

Adherence. The final stage is adherence to thenew recommendation. This involves the net effect oroutcome: whether or not patients consistently receiveimmunizations as recommended. Issues at work inthis stage include the environmental factors of parentagreement or demand for the immunization and prac-tice implementation policies, and informational fac-tors related to complexity of the new recommenda-

tion and compatibility with existing immunizationschedules and recommendations.

Parental agreement with new immunization rec-ommendations may be influenced by many informa-tion sources. Physician recommendations, friends,school or day care requirements, and even the newsmedia play a role in parents' impression of the safety,efficacy, and need for vaccines. Parental agreementis essential for a child to actually receive the vaccine.In some situations, parental demand for a vaccinerequired for school or day care will influence a physi-cian to offer that vaccine, regardless of whether heagrees with the recommendation.

Also affecting adherence is the manner in whicha new recommendation is implemented vis-a-vis ex-isting practice. For example, the addition of hepatitisB vaccine to the primary immunization series re-quired that infants receive three injections at one ormore well-child visit. Many parents, nurses, and evenphysicians initially balked at giving multiple injec-tions (41), the alternative, scheduling another ap-pointment, affected adherence to that recommenda-tion by delaying immunization receipt. Finally, thecomplexity of a new recommendation may prove con-fusing for physicians, office staff, and parents. Forexample, changes in the recommended age for ad-ministration of a vaccine may require the "relearn-ing" of standard routines by office nurses, revisionsto parent education materials, and an effective mech-anism to identify children who may be missed be-cause of the change.

LIMITATIONS TO THE MODEL

The model implies that all stages, from aware-ness to adherence, are equally significant and thatprogression through one stage is essential to reachthe next. However, some influences on physician be-havior may be so great as to cause stages to beskipped. For example, our empirical test of the model

Fig. 2. Effect of 95% progression through each stage in the awareness-to-adherence model.

238 Freed et al.

(13) identified several physicians who had adoptedthe universal hepatitis B vaccine recommendationeven though they did not agree with it. Analysis ofthe factors responsible found that both parental pres-sures and community practice norms resulted inthis phenomenon.

The model also does not demonstrate that forspecific recommendations or guidelines, certainstages may be more difficult to achieve than others.However, the model does allow for the identificationof the stage at which the progression to adherenceis arrested. Then, the factor responsible may be "di-agnosed" and "treated."

IMPLICATIONS FOR PRACTICE

The model presented here has an approach simi-lar to others in the literature that have looked atchanging patient behavior. Specifically, it is consis-tent with the recommendation from Prochaska thatinvestigators shift from the perception that behaviorchange is a specific action to one that involves stagesof progression (47). However, in contrast to individ-ual health-related behaviors, the risk of "relapse"or returning to the previous prerecommendation isminimal (48).

The concept of movement through sequentialstages involves certain mathematical consequencesregarding population-based estimates of eventual ad-herence to a guideline. For example, even if 95% ofphysicians achieve each step in the process, the resultis only an 80% adherence rate (Fig. 2)—a potentialproblem when striving to achieve herd immunity orother large-scale population coverage. Thus, if wide-spread adherence to any guideline follows a rationalseries of stages, it necessarily involves significant"buy-in" at each stage of the process. This placesconsiderable importance on effective guideline dis-semination.

This conceptual model of the process of imple-menting immunization recommendations has uses forboth research and practice. Although designed forimmunizations specifically, modification of individualfactors will allow its use in predicting or followingadoption of other guidelines. For organizations orgovernmental entities that craft and promulgate prac-tice guidelines, early recognition of the potential pit-falls along the way to adoption and adherence willfacilitate the planning process and suggest effectivestrategies for dissemination.

ACKNOWLEDGMENTS

Support of this project was provided by theAgency for Health Care Policy and Research(AHCPR R-01 HS 07286-0).

REFERENCES

1. Szilagyi PG, Rodewald LE, Humiston SG, et al. Immunizationpractices of pediatricians and family physicians in the UnitedStates. Pediatrics 1994;94:517-23.

2. Brown J, Melinkovich P, Gitterman B, Ricketts S. Missedopportunities in preventive pediatric health care. AJDC1993;147:1081-4.

3. Department of Health and Human Services. Healthy People2000. Washington, DC: Government Printing Office, 1990.

4. Lawrence RS. Diffusion of the U.S. Preventive Services TaskForce recommendations into practice. J Gen Intern Med1990;5:S99-S103.

5. Lomas H, Haynes RB. A taxonomy and critical review oftested strategies for the application of clinical practice recom-mendations: From "official" to "individual" clinical policy.Am J Prev Med 1988;4:77-94.

6. Kosecoff J, Kanouse DE, Rogers WH, McCloskey L, WinslowCM, Brook RH. Effects of the National Institutes of Healthconsensus development program on physician practice.JAMA 1987;258:2708-13.

7. Wood DL, Rosenthal P, Scarlata D. California pediatricians'knowledge of and response to recommendations for universalhepatitis B immunization. Arch Pediatr Adolesc Med1995;149:769-73.

8. Freed GL, Freeman VA, Clark SJ, Konrad TR, Pathman DE.A national assessment of family physician and pediatricianagreement with and adoption of universal hepatitis B immuni-zation. J Fam Pract 1996;42:587-92.

9. Freed GL, Bordley WC, Clark SJ, Konrad TR. Universalhepatitis B immunization of infants: Reactions of pediatriciansand family physicians over time. Pediatrics 1994;93:747-51.

10. Freed GL, Konrad TR, DeFriese GH, Lohr JA. Adoption of anew Haemophilius influenzae, type b vaccine recommendation.Am J Dis Child 1993;147:124-8.

11. Kraus DM, Campbell MM, Marcinak JF. Evaluation of uni-versal hepatitis B immunization practices of Illinois pediatri-cians. Arch Pediatr Adolesc Med 1994;148:936-42.

12. U.S. Department of Health and Human Services. Recom-mended childhood immunization schedule—United States,1995. MMWR 1995;44:RR-5.

13. Pathman DE, Konrad TR, Freed GL, Freeman VA, KochGG. The awareness-to-adherence model of the steps to clini-cal guideline compliance. Med Care 1996;34:873-89.

14. Freeman V, Freed GL, MaukopfA. Parental knowledge, atti-tudes, and demand regarding a vaccine to prevent varicella.Am J Prev Med, in press.

15. Clark SJ, Freed GL. Utilization of public immunization ser-vices after initiation of universal vaccine purchase. Arch PedAdolesc Med 1998;152:642-5.

16. Freed GL, Clark SJ, Pathman DE, Konrad TR, Biddle AK,Schectman RM. Impact of a new universal purchase vaccineprogram in North Carolina. Arch Pediatr Adolesc Med1997;151:1117-24.

17. Freed GL, Freeman VA, Mauskopf A. Enforcement of age-appropriate immunization laws. Am J Prev Med1998;14:118-21.

18. Freed GL, Kauf T, Freeman VA, Pathman DE, Konrad TR.

Adopting Immunization Recommendations 239

Vaccine-associated liability risk and provider immunizationpractices. Arch Pediatr Adolesc Med 1998;152:285-9.

19. Eisenberg JM. Sociologic influences on decision making byclinicians. Ann Intern Med 1979;90:957-64.

20. Freeborn DK, Baer D, Greenlick MR. Determinants of medi-cal care utilization: physician use of laboratory services. AmJ Public Health 1972:62:846-53.

21. Stolley PD, Becker MH, Lasagna L. The relationship betweenphysician characteristics and prescribing appropriateness.Med Care 1972;10:17-28.

22. Mick S (ed.). Innovation in the organization of health care:New insights in organization theory. San Francisco, CA: Jos-sey-Bass, 1991.

23. Williamson PM. The adoption of new drugs by doctors practic-ing in group and solo practice. Soc Sci Med 1975;9:233-6.

24. Carey T, Weiss K, Homer C. Prepaid versus traditional Medic-aid plans: Effects on preventive health care. J Clin Epide-miol 1990;43:1213-20.

25. Downs GW, Mohr LB. Toward a theory of innovation. Ad-ministration and Society 1979;10:379-408.

26. Pineault R. The effect of medical training factors on physicianutilization behavior. Med Care 1977; 15:51-67.

27. Lin LS, David MS. Physicians' orientation toward the legiti-macy of drug use and their preferred source of new druginformation. Soc Sci Med 1972;6:199.

28. Lin LS. Physician characteristics and attitudes toward legiti-mate use of psychotherapeutic drugs. J Health Soc Behav1971;12:132.

29. Coser RL. Authority and decision making in a hospital: acomparative analysis. Am Social Rev 1958;23:56-64.

30. Seeman M, Evans JW. Stratification and hospital care II. Theobjective criteria of performance. Am Social Rev1961;26:193-204.

31. Knafl K, Burkett G. Professional socialization in a surgicalspecialty: acquiring medical judgment. Soc Sci Med1975;9:397-404.

32. Haney CA. Psychosocial Factors Involved in Medical Deci-sion Making. In: Millon T (ed.), Medical behavioral science.Philadelphia, PA: W. B. Sanuders, 1975, pp. 420-32.

33. Burt RS. Social contagion and innovation: Cohesion versusstructured equivalence. Am J Social 1987;92:1287-1335.

34. Kaluzny A. Implementation of prevention and early detectionactivities: Selected organizational perspectives. In: Hibbard

H, (editors.), Proceedings from the Second Annual PrimaryCare Research Conference, 1991 (Jan), San Diego. Rockville,MD: DHHS/AHCPR, 1991.

35. Becker MH. Sociometric location and innovativeness: re-formulation and extension of the diffusion model. Am SocialRev 1970:35:267-82.

36. Mohr LB. Determinants of innovation in organizations. AmPol Sci Rev 1969:6:111-26.

37. Eisenberg JM. Physician utilization. The state of researchabout physicians' practice patterns. Med Care 1985:23:461-83.

38. Becker MH, Stoley PD, Lasagna L. Differential educationconcerning therapeutics and resultant physician prescribingpatterns. J Med Educ 1972:47:118-27.

39. Coieman JS, Katz E, Menzel H. MedicalInnovation. NewYork: Bobs-Merrill, 1966.

40. Tornatzky LG, Klein KJ. Innovation characteristics and inno-vation adoption-implementation: A meta analysis of findings.IEEE Transactions on Engineering Management1982;29:28-45.

41. Freed GL, Bordley WC, Clark SJ, Konrad TR. Reactions ofpediatricians to a new CDC recommendation for universalimmunization of infants with hepatitis B vaccine. Pediatrics1995:91:699-702.

42. Audet A-M, Greenfield S, Field M. Medical practice guide-lines: Current activities and future directions. Ann Int Med1990:113:709-14.

43. Pathman DE, Stevens CM, Freed GL, Jones BD, Konrad TR.Disseminating pediatric immunization recommendations: thephysician perspective. Amb Child Health, in press.

44. Freed GL, Freeman VA, Konrad TR, Pathman DE, ClarkSJ. Sources of immunization information for physicians. AmbChild Health 1997;2:329-37.

45. Mittman BS, Tonesk X, Jacobson PD. Implementing clinicalpractice guidelines: Social influence strategies and practitionerbehavior change. Qual Rev Bull 1992:18:413-22.

46. Karuza J, Calkins E, Feather J, Hershey CO, Katz L, MajeroniB. Enhancing physician adoption of practice guidelines. ArchIntern Med 1995;155:625-32.

47. Prochaska JO, Redding CA, Harlow LL, Rossi JS, VelicerWF. The transtheoretical model of change and HIV preven-tion: A review. Health Ed Q 1994:21:47-86.

48. Prochaska JO, Velicer WF, Guadagnoli E, Rossi JS, Diclem-ente CC. Patterns of change: Dynamic typology applied tosmoking cessation. Multivariate Behav Res 1991:26:83-107.