The interaction between policy and research in the implementation of server training

10
Addiction (1993) 88 (Supplement), 105S-113S The interaction between policy and research in the implementation of server training ERIC SINGLE Canadian Centre on Substance Abuse Policy and Research Unit, and University of Toronto, 100 College Street, Suite 207, Toronto, Canada M5G IL5 Abstract This paper describes the relationship between research and policy in the development of server training policy in Ontario. With the emergence of civil liability for the conduct of intoxicated patrons, there has been a growing movement toward the training of managers and servers in licensed establishment in the responsible service of alcohol. As part of a major government reform of alcohol regulations which has led to a new Liquor Licence Act, all persons who serve alcohol in licensed establishments in Ontario will be required to attend a server training course. The process leading to new alcohol regulations in Ontario is summarized with particular attention to the implementation of mandatory server training. The paper concludes with a discussion of the role of research in policy development. It is argued that the role of research in policy making is limited by an inherent conflict between the nature of scientific enquiry and the exigencies of policy development. This purpose of this paper is to describe the development of server training policy in Ontario and discuss the relationship between research and policy from the perspective of a public health researcher involved in the policy develop- ment process. The paper begins with a description of the emergence of civil liability for the conduct of intoxicated patrons and the con- sequent development of training programmes in the responsible service of alcohol. After extensive consultations by a special committee reviewing provincial alcohol regulations, the government has adopted a policy of requiring that all persons who serve alcohol in licensed establishments in Ontario will be required to attend a server train- ing course. The process leading to this policy is summarized, as well as issues of implementation. Finally, the role of research in development of server training policy is discussed. It will be argued that the research has had a limited role in policy development, in large part because the aims of scientific enquiry are not well suited to the requirements of successful policy advocacy. The emergence of civil liability and server training In Canada, persons who provide alcohol to others, either in licensed establishments or as social hosts, may be held liable for the damages and injuries caused by intoxicated patrons or guests. There are two major bases of civil liability in such situations. First, a server can be held liable as the provider of alcohol, as illustrated by the following case: Less than 15 minutes after finishing his last beer and leaving the Arlington House Hotel, Clayton Sharpe failed to negotiate a curve in the road. His car plunged over a steep em- 105S

Transcript of The interaction between policy and research in the implementation of server training

Addiction (1993) 88 (Supplement), 105S-113S

The interaction between policy and researchin the implementation of server training

ERIC SINGLE

Canadian Centre on Substance Abuse Policy and Research Unit, and University of Toronto,100 College Street, Suite 207, Toronto, Canada M5G IL5

AbstractThis paper describes the relationship between research and policy in the development of server training policyin Ontario. With the emergence of civil liability for the conduct of intoxicated patrons, there has been agrowing movement toward the training of managers and servers in licensed establishment in the responsibleservice of alcohol. As part of a major government reform of alcohol regulations which has led to a new LiquorLicence Act, all persons who serve alcohol in licensed establishments in Ontario will be required to attend aserver training course. The process leading to new alcohol regulations in Ontario is summarized withparticular attention to the implementation of mandatory server training. The paper concludes with adiscussion of the role of research in policy development. It is argued that the role of research in policy makingis limited by an inherent conflict between the nature of scientific enquiry and the exigencies of policydevelopment.

This purpose of this paper is to describe thedevelopment of server training policy in Ontarioand discuss the relationship between researchand policy from the perspective of a publichealth researcher involved in the policy develop-ment process. The paper begins with adescription of the emergence of civil liability forthe conduct of intoxicated patrons and the con-sequent development of training programmes inthe responsible service of alcohol. After extensiveconsultations by a special committee reviewingprovincial alcohol regulations, the governmenthas adopted a policy of requiring that all personswho serve alcohol in licensed establishments inOntario will be required to attend a server train-ing course. The process leading to this policy issummarized, as well as issues of implementation.Finally, the role of research in development ofserver training policy is discussed. It will beargued that the research has had a limited role in

policy development, in large part because theaims of scientific enquiry are not well suited tothe requirements of successful policy advocacy.

The emergence of civil liability and servertrainingIn Canada, persons who provide alcohol toothers, either in licensed establishments or associal hosts, may be held liable for the damagesand injuries caused by intoxicated patrons orguests. There are two major bases of civil liabilityin such situations. First, a server can be heldliable as the provider of alcohol, as illustrated bythe following case:

Less than 15 minutes after finishing his lastbeer and leaving the Arlington House Hotel,Clayton Sharpe failed to negotiate a curve inthe road. His car plunged over a steep em-

105S

106S Eric Single

bankment, rolling over repeatedly. Sharpe suf-fered only minor injuries, but his 16-year-oldpassenger, Andreas Schmidt, was rendered aquadriplegic'

This tragedy occurred in 1983 and receivedconsiderable media coverage. Schmidt brought a$13 million civil suit not only against the drunkdriver, Sharpe, but also against the ArlingtonHouse Hotel. The case was decided in favour ofSchmidt's claim, and the defendants were heldjointly and severally liable^ for $1 390 000. Thehotel was held responsible for 70% of Schmidt'slosses.̂ Taking into account pre- and post-judge-ment interest, legal fees and court costs, this casemay have cost the defendants and their insurersmore than $1 750 000. Schmidt sued the Arling-ton House Hotel because it provided the alcoholto Sharpe which made him intoxicated, andSharpe's intoxication was in turn a contributorycause of the accident in which Schmidt wasinjured.

There also exists a second and independentbasis of civil liability: in addition to being heldliable as alcohol providers, tavern owners andsocial hosts may be held liable as occupiers foralcohol-related injuries that occur on their prop-erty. An 'occupier' is defined as anyone who hascontrol over a premises and the power to admitor exclude others. In Jacobson v. Kinsmen Club ofNanaimo,'^ the defendant was a service clubwhich hosted a beer garden in a large arena. Apatron climbed an I-beam supporting the roof,dropped his pants and 'mooned' the capacitycrowd. Shortly afterward, another patron knownonly as 'sunshine' tried to imitate this feat."Sunshine", who was not as agile as his prede-cessor, lost his grip and fell onto the plaintiffJacobson, knocking him unconscious.("Sunshine", who was uninjured, left and wasnever heard from again.) The Kinsmen Club washeld liable for failing to take action to prevent theincident and permitting unsafe activities on thepremises.

Unfortunately, most occupier suits do not in-volve relatively harmless and humorous incidentssuch as this. In Stanton v. Twack (1982)' atavern was held liable when an intoxicated fe-male patron threw a glass into the plaintiffs face.In a more recent case involving a party held inLondon, Ontario, in 1986, an intoxicated guestbecame involved in a fight and was renderedparaplegic' Even though the host had not in-

vited the guest nor served alcohol to anyone, hewas liable for the safety of the premises and thecase was settled out of court for S700 000.

Until the 1970s, it was uncommon in Canadaand the US for individuals to be sued for theconduct of their intoxicated patrons, guests orcolleagues. However, during the past 15 yearsseveral factors have made such suits not onlypossible, but highly probable.' These factors in-clude:

(1) the expansion of the duties of positive oraffirmative action to control the conduct ofothers in a wide variety of situations;

(2) the introduction of mandatory testing ofbreath and blood samples in the drinking anddriving laws;

(3) the narrowing of the traditional defensesto civil liability;

(4) increases in the size of damage awards;(5) changes in public, government and judi-

cial attitudes.^

Thus, the emergence of civil liability for theconduct of the intoxicated has largely resultedfrom larger trends in tort law whereby personsare increasingly held liable for failure to takeaction when there is a foreseeable risk of harm toothers. It is not that the service of alcohol isbeing singled out as a special situation giving riseto liability, but rather that the service of alcoholis no longer being held to be exempt from thenormal principles governing civil liability.

While disconcerting to tavern owners and theirinsurers, the recent development of civil liabilityfor the conduct of the intoxicated offers newopportunities for the prevention of impaireddriving. Attempts to reduce the toll of deathsand injuries stemming from drunk driving havefocused almost exclusively on education of theindividual driver and the use criminal sanctionsagainst impaired driving. Such measures focuson individual drivers rather than situations whichgive rise to impaired driving incidence. Civil lawrepresents an entirely different approach. Al-though its main function is to provide for victimcompensation, the imposition of civil liability onpersons who provide alcohol represents a rela-tively new and potentially important mechanismfor the prevention of impaired driving and otherproblems associated with intoxication.

The increase in the number of law suitsagainst alcohol providers has caused insurancerates to rise dramatically. The threat of civil

Server training and the policy process 107S

liability, combined with increased communitypressure to reduce drunk driving, has "led theserving industry to seek new, creative ideas forestablishing reasonable guidelines for serving al-coholic beverages".'

In the past 10 years, hospitality organizations,government agencies and alcohol producers haveendorsed or developed server training pro-grammes in Canada.'" In 1983 a course onresponsible beverage service was developed bythe author for implementation in the hospitalitycurricula at two local community colleges inToronto. From 1984 to 1986 a number of servertraining programmes emerged in Canada, mainlysponsored by the hospitality industry, includingthose of private establishments and hospitalityorganizations such as the Canadian RestaurantAssociation." After 2 years of development, in1986 the Addiction Research Foundation (ARF)of Ontario introduced an extensive server inter-vention programme.'^ The ARF programme,which was formally endorsed by the Liquor Li-cence Board of Ontario (LLBO), consists of ahalf-day workshop for servers and a longer train-ing session for owners and managers. Theprogramme covers the following topics:

(1) alcohol and the law, including avideotaped presentation on civil liabilities;

(2) the health perspective, including the ef-fects of alcohol and factors which influenceimpairment;

(3) preventing intoxication, including how torecognize intoxication and diplomatically ceaseservice of alcohol;

(4) managing the intoxicated patron, includ-ing a discussion of safe transportationstrategies.''

The owners and managers course is lengthierbecause it involves more detailed examination ofmany issues and the development of a specificset of house policies aimed at preventing prob-lems in each participant's establishment.

Other Canadian programmes include thosedeveloped by provincial governments such asBritish Columbia, Manitoba and NewBrunswick.'" In 1986, the LLBO itself developeda half-day workshop for new licensees and trans-fers, which covers LLBO regulations andprocedures, as well as information on how torecognize and prevent intoxication from the ARFserver training programme.

Mandatory training under the new LiquorLicence Act of Ontario and the Role ofResearchIn 1986 a special Ontario Advisory Committeeon Liquor Regulation, was created to conduct anextensive review of alcohol regulation in theprovince." The Committee considered the issueof server training in detail. As the LLBO itselfhad developed a training programme, the Advi-sory Committee on Liquor Regulation was wellaware of the potential benefits of server training.During extensive public hearings held in 18 citiesthroughout the province, there was no opposi-tion to the concept of training programmes inresponsible beverage service. Earlier in 1986, thehospitality industry had strongly opposed a gov-ernment sponsored bill to introduce the sale ofbeer and wine in comer stores, which was even-tually defeated in the provincial Parliament. Thehospitality associations claimed that off-premisesale of alcohol should be placed in the hands oflicensed establishments rather than grocerystores on the grounds that they were experiencedin dealing with alcohol and the problems associ-ated with alcohol consumption. Given suchclaims of professionalization, it would have beendifficult for the hospitality industry to oppose theproposal that they be required to take a shortcourse in responsible beverage service.

In the absence of any opposition to the con-cept of mandatory server training, the AdvisoryCommittee on Liquor Regulation recommendedthat all servers in the province be required totake a training programme as a condition ofemployment in the hospitality industry. Manda-tory training was deemed necessary because itwas felt that otherwise only those establishmentswho are already reasonably responsible in theirserving practices would train their staff voluntar-ily. The Committee recommended further thatcourse content and standards be developed bythe LLBO, in collaboration with the ARF andhospitality interests. Cabinet approved this rec-ommendation and the LLBO is currentlydeveloping an implementation plan.

The recommendations of the Ontario Advi-sory Committee eventually led to a new LiquorLicence Act which the government proclaimedin October of 1990." The new law represented adramatic reversal in the direction of the policyreview. Although the Advisory Committee hadbeen mandated to sweep away what were viewedas outdated and unnecessary restrictions on alco-

108S Eric Single

hoi, the emphasis of the new Act is on socialresponsibility rather than liberalization of alcoholcontrols." Thus, the Ontario government de-cided to require server training in the context ofa general reform of provincial alcohol regulationswhich, on balance, involved tighter regulationover the service of alcohol rather than a liberal-ization of controls.

An assessment of the role of public healthresearch in the development of the new LiquorLicence Act concluded that the role of researchwas very limited." For most issues there waslittle research available (e.g. regarding the impactof changes in days or hours of operation ordetailed on-premise regulations), or the findingswere equivocal (e.g. the impact of alcohol adver-tising). However, one area of the new regulationswhich was influenced by research was servertraining, which provides an excellent example ofthe interaction between policy and research.

Research played a key role in server trainingpolicy in a number of ways. First, public healthresearchers have played an important role indeveloping and promoting server training pro-grammes.

Secondly, research has provided vital input toprogramme content. Recent experimental evi-dence indicating the reliability of visible signs ofintoxication as indicators of Blood Alcohol Con-centration" is but one example. Legal researchon the emergence of civil liability has been a vitalcomponent to the programmes.

Thirdly, evaluation research on server trainingprogrammes was a key factor in developing polit-ical support for mandatory training. Forexample, the National Steering Committee onImpaired Driving in Canada withheld supportfor a national conference on server interventionprogramming until it had first funded an evalua-tion study in Thunder Bay, Ontario.^" This studyand studies from the US provide qualified sup-port for the contention that server trainingprogrammes are effective in reducing intoxica-tion.^' In general, the evaluation studies indicatethat server training is effective in reducing patronintoxication without any major adverse economicconsequences to licensee profits.

While research has thus played an importantrole in developing and evaluating server training,it would be erroneous to claim that it had beenvital or even a necessary condition to the wide-spread development of server trainingprogrammes. It is doubtful whether there would

have been any significant development of servertraining programmes were it not for the emer-gence of civil liabilities of tavern owners, and thesupport of the hospitality industry and alcoholregulatory agencies for such training pro-grammes.

Training programmes must have the enlist-ment and support of the hospitality industry.Ultimately, the success of server training de-pends on its impact on the economic well-beingof drinking establishments. It is wrong to assumethat the hospitality industry is concerned solelywith maximizing alcohol sales. The principalconcern of licensed establishments is to maxi-mize profits, which is not necessarily the samething. The preliminary evidence fi^om servertraining evaluation studies indicates that formost establishments the loss of sales to intoxi-cated patrons is offset by increased sales of food,de-alcoholized or low-alcohol beverages, bylower insurance premiums and by the lower riskof civil liability. Research has thus helped con-vince the hospitality industry that it is goodbusiness to train serving staff, but no amount ofresearch would help if it were not in fact in theeconomic self-interest of the industry to adoptresponsible serving practices.

Mandatory training: implementation issuesIt is one thing to mandate server training, andquite another to implement this policy. Thereare myriad detailed issues involved in actuallydesigning and implementing a system to providetraining for up to 300 000 persons throughout aprovince the size of Ontario. At issue are whetherto develop a single programme or permit a freemarket to develop; whether to require training ofmanagers as well as servers; curriculum content;format and length of the training; certificationand training procedures for trainers; whether torequire testing and certification of trained serversand/or managers; provisions for continuous up-grading of programme content based on newinformation or legislation; provisions for trainingof language and/or cultural minorities; provisionsfor training in remote areas of the province; andmonitoring of the quality of training by licensingauthorities.

Although evaluation research has helped togenerate political support for the concept of

Server training and the policy process 109S

server training, research has offered little to in-form implementation decisions. Thusfar, servertraining evaluations have focused on whetherprogrammes are successful or not successful inmeeting goals, such as reducing the service ofalcohol to minors and intoxicated patrons. Therehas been virtually no research examining therelative merits of variations in programme con-tent, different formats or the cost effectiveness ofalternative programmes.

The LLBO began implementation by requir-ing server training for all licensed stadiums in theprovince in 1989 and all holders of a cateringlicense in the fall of 1990. The Board has alsoimposed, over the past 2 years, a requirement formandatory server training on certain licenseesthat have appeared before the Board on a varietyof disciplinary matters. In recent months, somenew licensees have also had this requirementimposed at the time of issuance of a new licence.

The implementation of mandatory training forthe remaining majority of licensed establish-ments is still in the planning stage. The LLBO iscurrently considering a proposal by the ARF tocollaborate in the development of a system forproviding mandatory server training throughoutthe province.̂ ^ The basic idea of the proposal isthat there should be one core curriculum or setof programme materials rather than many differ-ent programmes. The uniform core curriculumwould be under the control of the LLBO, butwith the major functions carried out by ARF,hospitality, community colleges and private orga-nizations.

The proposed system would involve threetypes of training programmes, one for owner/managers, one for servers and special trainingworkshops for trainers. Each would have a sin-gle, uniform curriculum, with a set of writtenback-up materials. The curriculum would beincorporated into training programmes whichwould be delivered by hospitality organizations,individual licensees, community colleges and pri-vate organizations. The content of the corecurriculum would be determined by the LLBOin consultation with an advisory committee con-sisting mainly of hospitality representatives. Thecurriculum would be developed and periodicallyupdated by the ARF, under a contractual ar-rangement with the LLBO. Trainers would bemonitored and trainees certified by the ARF,while LLBO inspectors would be responsible forchecking that serving staff are certified as having

been trained. Training fees would be used torecover costs for the programme delivery and formonitoring. If the need is not filled by the ordi-nary market mechanisms, special arrangementswill be made for the delivery of training in lan-guages other than English and for delivery oftraining in remote areas of the province.

The ARF proposal argues that this deliverysystem offers a number of advantages comparedwith alternative methods of providing for man-datory server training. First and foremost, itassures quality control. The alternative 'freemarket' system for providing mandatory trainingnecessarily gives a competitive advantage to thelower quality programmes which spend less oncurriculum development and quality control.Secondly, the proposed system involves all majorinterest groups in an appropriate role. TheLLBO retains decision making authority on allaspects of server training. The curriculum devel-opment, training and monitoring of trainers, andcertification of trainees would be largelyconducted by ARF. The actual training wouldbe delivered by hospitality groups, licensees,community colleges and private organizations.Thirdly, the proposed system would involve aminimum of LLBO staff time and resources asmost of the operational functions would behandled by other organizations. The LLBOwould necessarily be involved in an inspectionsfunction, but otherwise it is likely that it couldfulfil its role with only one or two full-time oreven part-time staff. Fourthly, the proposed sys-tem allows the LLBO to integrate mandatorytraining into more general strategies to commu-nicate more effectively with licensees. Finally,the proposed system incorporates processes forthe continual updating of information in thetraining programme.

The LLBO has yet to decide on this proposal.Although the major hospitality groups initiallyendorsed the proposal, some now oppose certainaspects of the proposed plan. The ARFprogramme is based on a system of geographi-cally defined franchises and hospitalityorganizations are seeking the right to give theprogramme to their members anywhere in theprovince. They are also concerned that fees vnWbe high. However, negotiations are under wayand it is likely that some compromise will bereached. Hopefully, in the near future the LLBOwill armounce guidelines for mandatory servertraining.

11 OS Eric Single

Conclusions: the inherent conflict betweenscience and the exigencies of policydevelopmentThus, even in the relatively successful area ofserver training, research has not been pivotal andin general it must be concluded that publichealth research has generally played a very lim-ited role in alcohol policy development. Part ofthe reason for this is limited access to policymaking. Scientists are rarely involved at thehigher levels of policy deliberations.^' However,in most situations, there is some degree of ac-cess, typically via the presentation of a policydocument. Part of the reason why scientists donot take greater advantage of opportunities toprovide input into public policy is that they donot generally receive peer recognition or careeradvancement for spending a good deal of time inpolicy development.

The lack of rewards and lack of access may bemore consequences than causes of ineffectivepolicy advocacy on the part of researchers. Themajor reason for the lack of greater influenceconcerns an inherent conflict between the natureof scientific enquiry and the exigencies of policydevelopment. The aim of science can be de-scribed as being to discover and understandabsolute truth. Science aims at the progressivedevelopment of an immutable body of knowl-edge. Policy, on the other hand, is generallyconcerned with what can be done accomplishednow, with limited knowledge, time and re-sources. This difference is perhaps best illus-trated by considering the essential characteristicsof successful policy documents. A policy advo-cacy document is more likely to influence thepolicy decision if the following conditions aremet.

/. Credibility of the sourceThe greater the extent to which the informa-tion presented in a policy document isviewed as objective, accurate and fair, thegreater its influence. Credibility is enhancedby the acknowledgement of disconfirmingevidence, the consideration of alternativepolicy options, the ability to cite well-de-signed evaluation studies, the ability to pointto the successfiil use of policy instruments inother jurisdictions, and the lack of obviouspersonal gain or self-interest on the part ofthose advocating a particular policy. This isclearly one of the most important strengths

of public health and other scientific presen-tation on policy issues, which tend to havehigh credibility with policy makers.

Credibility, however, is fragile: once an in-dividual or organization is viewed as havingpresented a biased policy recommendation,future representations will be viewed withscepticism. Personally, I am not willing tostake my individual credibility on issues un-less there is strong scientific justification fora particular position. Thus, for example,even though I would be happy to see a banof all alcohol advertising, I did not advocatethis position during the Advisory Committeedeliberations. There was simply insufficientevidence on the impact of alcohol advertis-ing to risk my credibility on this issue. Onthe other hand, backed by strong evaluationstudies indicating a significant impact, Istrongly advocated the policy regardingserver training. Influence is a coin which ishard to obtain, so it is important to spend iteffectively.

2. A broad policy perspective

Although public health policy documentstend to be viewed as credible, they also tendto be very narrowly focused on health out-comes. It is inevitable that such research willonly play a minor role in policy delibera-tions. Few policy issues can be reduced toempirical questions focusing on publichealth outcomes alone, whereby specificallydesigned studies can provide the decisiveinput.

The narrow policy perspective typical ofmost scientifically based policy documentsentails relatively low risk: it does not requireone to venture beyond one's area of exper-tise or risk personal credibility. However anarrow policy perspective is also easily dis-missed in policy debates, particularly at theCabinet level where short-term economicconsiderations tend to be given greaterweight than long-term public health consid-erations.

The adoption of a broad policy perspec-tive is therefore vital if one wishes to beeffective in the policy arena. Considerationsof scientific accuracy and the progressivedevelopment of knowledge invariably lead tothe reward of specialization in research.Effectiveness in policy debates requires theability to offer specialized knowledge and

Server training and the policy process 111S

keep it well-grounded in a broad perspectivewhich takes political, economic, cultural andhistorical factors into account. This is a rarequality among scientists.

3. Receiving broad supportA corollary to the principle that a broadpolicy perspective tends to be more effectiveis that policy positions which are broad-based, with support from many groups inthe community, are more likely to be wellreceived by policy makers. Server trainingclearly enjoys widespread support, includingregulatory agencies, public health interests,citizen groups, insurance companies andeven the alcohol industry. Other publichealth orientated policies, such as increasedtaxation or a ban on advertising, generatestiff resistance from the alcohol industrywhich policy makers must take into consid-eration.

In order to develop a broad base of sup-port, it is frequently necessary tocompromise with alternative interests. Forexample, server training programmes havehad to eliminate or modify public healthgoals to be acceptable and practical for thehospitality industry. Scientists are not neces-sarily inept at compromise and thegeneration of support from other interests,but to do so they must be able to communi-cate effectively in non-technical fashion.More importantly, the development of sup-port for policy requires a sense of give andtake. If I could not convince the members ofthe Advisory Committee of the wisdom ofmy position on a particular issue, then I hadto acknowledge that perhaps my ownposition required reassessment.

4. Communicated in a clear and concise fashionPolicy documents must be communicatedclearly and succinctly. They must, first, belegible to their intended audience. Technicalconcepts and scientific jargon may enhancethe ability of scientists to communicate withother specialists, but they diminish effective-ness in policy advocacy.

Furthermore, policy documents must beconcise. Government ministers and theirstaff must continuously consume an enor-mous volume of information. The briefer apolicy document, the more likely that it willbe read and have some infiuence on a policydecision. Brevity can be very difficult with

regard to the presentation of public healtharguments. Important qualifiers may beomitted, and even then it may be difficult topresent a succinct yet complete exposition ofpublic health oriented policy options.

5. PromptLast but not least, policy documents must beprepared promptly. It is virtually useless topresent a policy document, however con-vincing and well supported, if it is receivedafter a policy decision has been made. Ofrena policy document is only effective if it canbe prepared within a matter of days or evenhours.

Scientific enquiry is extremely conscious ofcredibility, but scientists frequently lack, or evennegatively evaluate, the other essential character-istics of successful policy advocacy. Sciencerewards specialization rather than adopting abroad perspective on issues. It requires the de-velopment of technical language and highlyspecialized knowledge which can inhibit the abil-ity to communicate with lay persons and enlistsupport. The length of scientific writing tends tobe dictated mainly by content, and scientificarguments can rarely be made succinct withoutreliance on jargon or technical concepts whichonly other scientists are likely to understand.The development of science is not generallyseem as a short-term activity. Scientists are gen-erally expected to take as long as necessary toconduct an investigation, and lines of researchtypically evolve over years. They are thereforeunaccustomed to the short deadlines which typi-cally apply to policy development.

For all of these reasons, public health researchis not well suited to policy advocacy. Perhaps themajor reason why scientists are not comfortablein policy arenas, however, is that effectiveness inpolicy debates entails personal risk and scientistsare, by both nature and training, cautious indi-viduals. Once the public health professionaltakes into account factors beyond one's area ofexpertise, one is forced to shed the cloak ofscientific objectivity and respectability. There isan expression among policy makers that experts"should be on tap, not on top." Scientists whoenters the policy arena can no longer viably claimto be detached and objective.

The risks involved in policy advocacy are high.It is, of course, a matter for each individual todecide whether the risks are worth the potential

112S Eric Single

gains when one takes a stand on a particularissue. Although the potential benefits are enor-mous for many health issues, I am notnecessarily arguing that public health profession-als should play a greater role in policy debates.As indicated earlier, I am personally very cau-tious regarding which issues I choose to adopt anadvocacy role. I am merely pointing out that it ishypocrisy to complain that our work is not givengreater consideration if we as individuals are notwilling to step outside of the comfortable mantleof scientific respectability. In so far as possible,policy should be developed using scientific meth-ods, based on empirical data which have beenscientifically collected and analyzed. But ulti-mately alcohol policy issues are social issues,involving normative decisions, which are there-fore best resolved via the political process despiteits shortcomings.

Notes and References1. SOLOMON, R., BLOAKE, B. & GLEASON, M . (1985)

One for the road: a tavern owner's liability as aprovider of alcohol, in: SINGLE, E. & STORM, T .(Eds) Public Drinking and Public Policy, p. 257(Toronto, Addiction Research Foundation).

2. The principle of "joint and several liability" ren-ders licensed establishments particularlyvulrierable to large damage awards in civil suits.By virtue of provincial statute (e.g. the OntarioNegligence Act), in negligence suits a defendant isheld jointly and severally liable with all of hisnegligent co-defendants. A successful plaintiff maytherefore collect the entire amount awarded tohim or her from any one of the defendants heldliable even if the particular defendant was onlydeemed partially responsible. Thus, potential de-fendants with the most assets, such as bars andtaverns, tend to be chosen by plaintiffs as de-fendants in negligence suits and frequently mustpay for the lion's share of damages even whendeemed only partially responsible.

3. Schmidt v. Sharpe et al. (1983), 27 CCLT 1(Ontario Supreme Coun).

4. Jacobson v. Kinsmen Club of Nanaimo (1976) 71DUR (3d) 227 (British Columbia SupremeCourt).

5. Cited in SOLOMON, R. & USPRICH, S. (1990) Civilliability for the conduct of the intoxicated acrossCanada, in: SINGLE, E. (Ed.) Server Intervention inCanada, pp. 46-96 (Ottawa, Health and WelfareCanada).

6. Cited in CANADIAN CENTRE ON SUBSTANCE ABUSE(1991) Being sued can ruin a good party (Ottawa,Canadian Centre on Substance Abuse).

7. Canadian law governing the liability of alcoholproviders consists mainly of common law (prece-dents established from the accumulation of case

law), as well as statutory provision of liability inManitoba, Nova Scotia, Ontario and the North-west Territories. Common law is applicable to allof Canada other than Quebec, which is governedby civil law. However, the provisions of liabilityembodied in the Quebec CivU Code are essentiallythe same as those in common law. See SOLOMON,R. & USPRICH, S. (1990) Civil liability for theconduct of the intoxicated across Canada, in:SINGLE, E. (Ed.) Server Intervention in Canada, pp.46-96 (Ottawa, Health and Welfare Canada).

8. SOLOMON, R. & USPRICH, S. (1990) Civil liabilityfor the conduct of the intoxicated across Canadain: SINGLE, E. (Ed.) Server Intervention in Canada,pp. 46-96 (Ottawa, Health and Welfare Canada).Furthermore, it is unlikely that government vsdllintercede to arrest this trend. The Ontario Advi-sory Committee on Liquor Regulation examinedthe civil liability issue in detail and recommendeda strong, new, comprehensive, fault-based statu-tory provision of liability. See ONTARIO ADVISORYCOMMITTEE ON LJQUOR REGULATION (1987) Re-port of the Advisory Committee on LiquorRegulation/Rapport du comite consultif sur la regie-mentation des alcools, Ch. 9 (Toronto: Queen'sPrinter). The Ontario government has endorsedthese recommendations and action on their imple-mentation is pending. It is noteworthy thatalternative recommendations which might havediminished the liability of tavern owners and otherproviders of alcohol were explicitly rejected by theOntario Advisory Committee on Liquor Regula-tion. Thus, for example, the Committeeconsidered but decided against recommendationsto limit the size of damage awards, require manda-tory insurance, provide special notice provisions incases involving alcohol providers, or provide aspecial exemption in alcohol cases from the princi-ple of "joint and several liability".

9. MOSHER, J. (1984) Server intervention: presentstatus and future prospects, paper presented at theResearch Workshop on Alcohol and the DrinkingDriver, Bethesda, Maryland, p. 3.

10. SINGLE, E. (1989) Server intervention: a new ap-proach to the prevention of impaired driving.Health Education Research, 5, pp. 237-245.

11. REYNOLDS, J. (1985) The Responsible Service ofAlcoholic Beverages: A Guide For the HospitalityIndustry (Toronto Canadian Restaurant Associa-tion).

12. SIMPSON, R., STANGHETTA, P., BRUNET, S.,SINGLE, E., SOLOMON, R. & VAN DE BCLUET, W.(1986) A Guide to the Responsible Service of Alcohol(Toronto, Addiction Research Foundation).

13. BRUNET, S. (1990) Key components of servertraining, in: SINGLE, E. (Ed.) Server Intervention inCanada, pp. 22-27 (Ottawa, Health and WelfareCanada).

14. See ATKINSON, G. (1990) The development ofresponsible service program in the province ofNew Brunswick, in: SINGLE, E. (Ed.) Server Inter-vention in Canada (Ottawa, Health and WelfareCanada), pp. 28-32; CARLSON, J. (1990) BritishColumbia's server and licensee training program,in: SINGLE, E. (Ed.) Server Intervention in Canada

Server training and the policy process 113S

(Ottawa, Health and Welfare Canada), pp, 33-38;and MANITOBA LIQUOR CONTROL COMMISSION(1987) One too many: recognizing the signs ofintoxication (Winnipeg, Manitoba Liquor ControlCommission),

15, See SINGLE, E. & TOCHER, B. (1992) Legislatingresponsible alcohol service: an inside view of thenew Liquor Licence Act of Ontario, British Journalof Addiction, 1992 (in press). This article describeshow the Advisory Committee framed the issuesand arrived at 73 recommendations on issues suchas licence classification, the control of special per-mit functions, days and hours of operation, legaldrinking age, alcohol advertising and the civilliabilities of alcohol providers. The key actors andconsiderations are discussed, as well as the subse-quent Cabinet action on the report and theimplementation process currently underway. Thepaper concludes with an assessment from both anadministrative and a public health perspective,

16, There were several reasons why it took three anda half years to pass new legislation based on theAdvisory Committee repon. The Advisory Com-mittee reported report contained 73recommendations, and Ministry staff expected arange of policy options to be investigated for eachrecommendation. More than 200 policy optionswere researched and developed for the Minister'sconsideration. Twice there were changes in Mini-sters, necessitating reconsideration of many issues.Extensive consultations were required with otherministries, and the actual drafting of the rathercomplex piece legislation took approximately ayear,

17, Although the licensing system has been simplifiedand certain relatively minor regulations have beenrepealed (e,g, interdiction orders and the require-ment for a hospitality license for 'boardroombars'), the major changes involve the imposition ofnew and stricter rules concerning the sale andservice of alcohol. Mandatory server training is butone example of this. All licensees are now requiredto provide food. The civil liabilities of providers ofalcohol remain intact, and may eventually becodified in a new statute, no longer subject tochanging judicial interpretation of common lawdoctrines. Special occasion permit functions arenow better controlled in a variety of ways. Adver-tising regulations have been strengthened, andpublic service messages on the hazards of alcohol

consumption have increased substantially. TheLLBO authority to enforce the drinking age inlicensed establishments has been strengthened. Itis also noteworthy that several significant liberal-ization measures were considered but rejected. Inparticular, the minimum drinking age of 19 wasnot lowered to 18, and there was no major changeto the hours of operations for licensed establish-ments,

18, SINGLE, E. & TOCHER, B. (1992) Legislating re-sponsible alcohol service: an inside view of thenew Liquor Licence Act of Ontario, British Journalof Addiction, 87, pp, 1433-1443,

19, TEPUN, L & LuTZ, G. (1985) Measuring alcoholintoxication: the development, reliability andvalidity of an observational instrument. Journal ofStudies on Alcohol, 46, pp, 459-466,

20, The proceedings for tiiis conference, which washeld in Toronto in 1988, are reported inSINGLE, E, (Ed,) Server Intervention in Canada(Ottawa, Health and Welfare Canada, 1989), TheThunder Bay evaluation study is presented by L,Gliksman in the conference proceedings,

21, For a review of the Canadian and the five U,S,studies, see SINGLE, E. (1989) Server intervention:a new approach to the prevention of impaireddriving. Health Education Research, 5, pp, 237-245, and SINGLE, E. (1991) Trends in civil liabilityand server intervention in Canada, paper pre-sented at the Perspectives for Change conference,Wellington, New Zealand,

22, SINGLE, E. (1990) A uniform LLBO training pro-gramme—a proposal to the LLBO and theworking committee on the implementation ofmandatory server training in Ontario (Toronto:Addiction Research Foundation), Although theauthor left ARF to join the Canadian Centre onSubstance Abuse shortly after developing this pro-posal, he continues to have close collaborative tiesto both the LLBO and ARF, The details of theARF proposal have already been made public byARF,

23, The author's presence on the Ontario AdvisoryCommittee on Liquor Regulation is an exceptionto the normal situation in which researchers areexcluded from such bodies. And even in this situ-ation he did not have a vote. It is likely, however,that the mere presence of someone representingpublic health interests infiuenced the direction ofthe policy review.