UBC Faculty Research and Publications

Tool, weapon, or white elephant? A realist analysis of the five phases of a twenty-year programme of… Spiegel, Jerry M; Lockhart, Karen; Dyck, Carmen; Wilson, Andrea; O’Hara, Lyndsay; Yassi, Annalee Aug 6, 2012

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-12911_2011_Article_577.pdf [ 368.83kB ]
JSON: 52383-1.0223302.json
JSON-LD: 52383-1.0223302-ld.json
RDF/XML (Pretty): 52383-1.0223302-rdf.xml
RDF/JSON: 52383-1.0223302-rdf.json
Turtle: 52383-1.0223302-turtle.txt
N-Triples: 52383-1.0223302-rdf-ntriples.txt
Original Record: 52383-1.0223302-source.json
Full Text

Full Text

RESEARCH ARTICLE Open AccessTool, weapon, or white elephant? A realistanalysis of the five phases of a twenty-yearprogramme of occupational health informationsystem implementation in the health sectorJerry M Spiegel*, Karen Lockhart, Carmen Dyck, Andrea Wilson, Lyndsay O’Hara and Annalee YassiAbstractBackground: Although information systems (IS) have been extensively applied in the health sector worldwide, fewinitiatives have addressed the health and safety of health workers, a group acknowledged to be at high risk ofinjury and illness, as well as in great shortage globally, particularly in low and middle-income countries.Methods: Adapting a context-mechanism-outcome case study design, we analyze our team’s own experience overtwo decades to address this gap: in two different Canadian provinces; and two distinct South African settings.Applying a realist analysis within an adapted structuration theory framing sensitive to power relations, we explorecontextual (socio-political and technological) characteristics and mechanisms affecting outcomes at micro, mesoand macro levels.Results: Technological limitations hindered IS usefulness in the initial Canadian locale, while staffing inadequaciesamid pronounced power imbalances affecting governance restricted IS usefulness in the subsequent Canadianapplication. Implementation in South Africa highlighted the special care needed to address power dynamicsregarding both worker-employer relations (relevant to all occupational health settings) and North–south imbalances(common to all international interactions). Researchers, managers and front-line workers all view IS implementationdifferently; relationships amongst the workplace parties and between community and academic partners have beenpivotal in determining outcome in all circumstances. Capacity building and applying creative commons and opensource solutions are showing promise, as is international collaboration.Conclusions: There is worldwide consensus on the need for IS use to protect the health workforce. However, ISimplementation is a resource-intensive undertaking; regardless of how carefully designed the software, contextualfactors and the mechanisms adopted to address these are critical to mitigate threats and achieve outcomes ofinterest to all parties. Issues specific to IS development, including technological support and software licensingmodels, can also affect outcome and sustainability – especially in the North–south context. Careful attention mustbe given to power relations between the various stakeholders at macro, meso and micro levels whenimplementing IS. North–South-South collaborations should be encouraged. Governance as well as technologicalissues are crucial determinants of IS application, and ultimately whether the system is seen as a tool, weapon, orwhite elephant by the various involved parties.(Continued on next page)* Correspondence: jerry.spiegel@ubc.caGlobal Health Research Program (GHRP), School of Population and PublicHealth, University of British Columbia (UBC), Vancouver BC V6T 1Z3, Canada© 2012 Spiegel et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the CreativeCommons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, andreproduction in any medium, provided the original work is properly cited.Spiegel et al. BMC Medical Informatics and Decision Making 2012, 12:84http://www.biomedcentral.com/1472-6947/12/84(Continued from previous page)You may call me a fool, But was there a rule The weapon should be turned into a tool? And what do we see? Thefirst tool I step on Turned into a weapon. - Robert FrostWhite (albino) elephants were regarded as holy in ancient times in Thailand and other Asian countries. Keeping awhite elephant was a very expensive undertaking, since the owner had to provide the elephant with special foodand provide access for people who wanted to worship it. If a Thai King became dissatisfied with a subordinate, hewould give him a white elephant. The gift would, in most cases, ruin the recipient. - The Phrase FinderBackgroundDespite expenditure of billions of dollars worldwide ininformation system (IS) applications, controversy per-sists concerning what this extensive investment hasachieved [1-4]. While IS implementation must ultimatelybe assessed with regard to applications in specific socialcontexts and not in the abstract, this orientation haslargely been neglected [5,6]. To address this challenge,the structurationist theoretical orientation [7,8] empha-sizes that each IS configuration decision “is not merelytechnical, but social and political, affecting end-users’practices” [8]. However, consideration of micro-, meso-and macro-level influences on relevant outcomes hasreceived limited consideration [9] and the structuration-ist framing itself has been the subject of considerablecritique, including shortcomings in adequately apprecia-ting the nuances of contextual factors [10]. Furthermore,while Myers & Klein observed that IS applications areincreasingly addressing “social issues such as freedom,power, social control, and values with respect to the de-velopment, use, and impact of information technology”[11], the direct engagement in IS implementation ofthose whose health is directly in question still remainslargely ignored. This contrasts with the active debatethat has occurred regarding how geographic informationsystems have been implemented in circumstances whereresidents of affected communities have felt victimizedwhen their involvement has been marginalized [12,13].Among those cognizant of the seminal importance ofcontext, the power implications related to how a “surveil-lance gaze” is applied have stimulated much reflection[14]. Critical theorists have given explicit considerationto how resistance to intended IS application may bemanifest, such as by health professionals who attempt toadapt its application to their own perceived interests[14] or by those who respond to being observed by seek-ing to disrupt smooth functioning and control [15].However, in addition to ambiguities rooted in whetheran IS is perceived from the perspective of specificgroups as either contributing value (as a “tool”) or threa-tening particular interests (as a “weapon”), IS technologytransfer is always fraught with the danger of being a“white elephant” that confers limited benefit and drainsresources.To mitigate obstacles to access restrictions that can beprovoked by proprietary ownership of intellectual pro-perty (not unlike the debates regarding generic drugs asan option to reliance on patent protection of pharma-ceuticals), IS innovations have also triggered develop-ment of alternative licensing and knowledge-sharingorientations. These include Creative Commons licensing,in which software products are made available with therequirement of attribution, non-commercialization, andeither no derivatives allowed, or a share-alike model, inwhich modifications and new developments are thenshared with “the commons” [16]. This is potentially animportant strategy in mitigating power dominance.“Open Source Software” (OSS), which operates on thisprinciple, nevertheless still requires a dedicated inter-national group of skilled developers, or at least a well-resourced passionate host to maintain the system [17].Moreover, Hertel et al. note that even in the OSS com-munity, contributors’ motivations to OSS may primarilybe to improve their own software (i.e. learning oppor-tunities) and participation (i.e. as part of a large team),and still not necessarily involve ultimate users of thesoftware [17].Another important contextual factor, one that reflectsNorth–South power imbalances in particular, is the“digital divide”, which refers to the wide disparities thatcurrently exist between high income countries (HICs)that have developed IS solutions in comparison to lowand middle income countries (LMICs) where capacitiesfor developing and applying such undertakings may bequite limited [18,19]. Studies on information technologyin LMICs [20-23] repeatedly emphasize the need to bet-ter understand how to best introduce health IS fordecision-making in these settings – which is of particu-lar relevance given the disproportionately greater globaldisease burden that occurs here.Ultimately, as Guba & Lincoln observed, stakeholders’assessment of an intervention’s worth largely depends onhow the causes of underlying problems are perceived[24]. In this regard, those who design a particular inter-vention, or govern its development and use, typically dothis with a different rationale or perspective than thosewho are affected; and designers, decision-makers, anddeliverers might themselves maintain distinct norms andSpiegel et al. BMC Medical Informatics and Decision Making 2012, 12:84 Page 2 of 15http://www.biomedcentral.com/1472-6947/12/84values [25]. Accordingly, the power relations embeddedin settings where information systems are used must beaddressed [26].Finally, while pronounced “management–employee”power differences are manifest in all workplace settings,the right to be informed and involved in addressing rele-vant health determinants, especially through bipartitemanagement-worker occupational health and safety(OHS) committees, is guaranteed by statute in a largenumber of jurisdictions worldwide [27]. However, al-though such “agents” (as we refer to this unit in Figure 1below) exist in Canada [28], the United States [29], theUnited Kingdom [30], and Australia [31] as well as inEuropean [32] and Asian countries [33] in addition toLMICs (including South Africa [34]), it must beacknowledged that they are not universal, and tend to beweaker in non-union environments and particular sub-sectors such as homecare within the healthcare sector.Figure 1 presents the basic relationships that we setout to explore in scrutinizing “IS implementation rele-vant” power relations in workplace settings. The theo-retical orientation that we apply can be characterizedas an “adapted structurationist” model, allowing forgreater exploration of power asymmetries within theworkplace as well as at a global level. Adopting thisapproach, our article examines how contextual factors,and specifically power relations, can play a fundamen-tal role in affecting IS implementation. In this sense,while applying a structurationist orientation to the“dualistic” influence of structure and social practices,we investigate the particular sensitivities to power rela-tions inherent in this dialectic, as they relate to thevarious challenges highlighted above and integratedwithin Figure 1.The health sector workplaceThe health sector constitutes one of the largest sourcesof employment worldwide [35], with acute shortages ofhealth workers a serious concern, especially in LMICsettings [36]a. Healthcare systems worldwide are notonly plagued by difficulties in recruitment and retention,but also biological, chemical and physical occupationalhealth hazards as well as ergonomic hazards and psycho-social factors that lead to a high risk of injuries, illnessand stress [35]. Nevertheless, despite the increasing useof online information systems in occupational healthgenerally [37], and the existence of OHS committeesthat crave information for decision-making [38], anddespite the fact that the health sector has a culture inwhich health-related surveillance is recognized as im-portant, the use of IS has been rarely applied to improvethe sector’s own work environments. The impact ofglobalization can, at least partly, explain this inattentionto the well-being of the healthcare workforce – withincreased casualization of work, subcontracting servicesand weakening of healthcare unions [39]. These factorsalso create a disincentive to reporting of hazards, letalone addressing these. As such, underreporting hasFigure 1 Adapted structuration model guiding analysis of Information System (IS) use for Occupational Safety and Health (OSH).Spiegel et al. BMC Medical Informatics and Decision Making 2012, 12:84 Page 3 of 15http://www.biomedcentral.com/1472-6947/12/84been identified as a serious problem hindering accuratesurveillance and appropriate action [40,41].Accordingly, examining situations in which IS inno-vations have indeed been implemented in the health-care workplace provides an excellent opportunity toanalyze contextual influences in circumstances whereother factors (knowledge, technical capacity, focus, re-cognition of rights) could support successful imple-mentation. Acknowledging the power asymmetries inthis setting, we particularly examine how an IS couldserve (or be perceived) as a tool towards promotingthe health of the healthcare workforce, or a weapon inservice of other objectives. The objective of this studywas to draw lessons from our own experience im-plementing information systems to address the healthof healthcare workers in various contexts over twodecades.MethodsWe analyzed our team’s experiences over two decades inchronological order, first: i) in Canada – the provinces ofa. Manitoba and b. British Columbia; and then ii) inSouth Africa – a. the province of Free State and b. theNational Health Laboratory System (NHLS). We also re-flect upon recent further international collaboration thathas been developing to address the challenges identified.Given the prominence of the Canadian and South Africanexperience in informing international efforts currentlyunderway [42,43], and the leadership role these activitiesare playing within the World Health Organization (WHO)and International Commission on Occupational Health(ICOH)’s Scientific Committee on Occupational Healthfor Health Care Workers [44], an in-depth analysis ofthese experiences is especially warranted.For this analysis, the context-mechanism-outcomemethod [45] was applied, in which we describe thesocio-political context of each of the four initiatives, de-lineating the mechanisms employed to achieve success,and analyzing the outcome. This is the general approachused in “realist reviews” [45] - a strategy for synthesizingresearch that has an explanatory rather than judgmentalfocus. In realist evaluation, to infer a causal outcome (O)between two events, one needs to understand the under-lying mechanism (M) connecting them and the context(C) in which the relationship occurs, with the basicevaluative question of ‘what works?’ replaced by ‘what isit about this program that works for whom in what cir-cumstances?’ [46]. Mechanisms, moreover, are sensitiveto variations in context, as well as to the operation ofother mechanisms in a particular context [47]. We expli-citly apply a structurationist theory approach sensitive topower relations in each of these contexts.Several sources of information were used for con-structing the C-M-O (Context – Mechanism – Outcome)analysis, as is summarized by Table 1. This encom-passed: a) articles published about the context, theimplementation of the system or from using data gener-ated by these initiatives; as well as from b) surveys ofusers, c) key informant interviews and d) our own obser-vations as researchers either directly involved in design-ing and implementing the systems in question (authorAY), using the data produced (authors JS, AY, KL andLO), or studying the implementation of the systems inquestion (authors CD, AW and JS), as outlined further inTable 1.Rather than detailing specific experiences, we focus ondistinct characteristics associated with variations in thedifferent applications that our team has pursued overthe years, in order to especially assess the influence ofTable 1 Summary of IS implementation review data sources1. 2. 3. 4.Setting Winnipeg, Manitoba, Canada British Columbia, Canada Free State, South Africa Nationally, South AfricaIS established at a largehealth facility with bipartiteOHS committeeIS established at provinciallevel governed bybipartite boardIS built & piloted athospitals in provincewith bipartite oversightIS applied in a multi-site(349 labs) national institutionObservation Period 1986 - 1999 2000 - 2011 2007 - 2012 2010 - 2012Data Sourcesa. Articles revieweda 18 peer review articles 23 peer review articles 4 peer review articles 1 peer review abstractb. User surveysb Captured by participantobservationManager & front lineworker surveysPilot study interviews;survey conductedPreliminary surveyc. Key informantinterviewscCaptured by participantobservationInterviews of managersfrom 2 health regions &the provincial OHS regulatorInterviews of 2 OHSchampions &other managersInterviews of IS &OHS managersd. Participant-observeranalysesdResearcher leadingdesign & use;regulator perspectiveResearcher leadingdesign & use;research managerResearcher co-leading design &use; research manager &coordinator; evaluatorResearcher co-leadingdesign & use; evaluatorNotes: Details associated with notes a, b, c and d are provided in the Appendix.Spiegel et al. BMC Medical Informatics and Decision Making 2012, 12:84 Page 4 of 15http://www.biomedcentral.com/1472-6947/12/84contextual factors, and specifically the power relationswithin these. As was the approach of Porter in studyingpower relations between nurses and physicians inhealthcare [48], the method of analysis we used couldbe called analytical induction, whereby tentative hy-potheses about the power relations in the work envi-ronment were constantly refined, altered or abandonedin the light of the data collected. As noted by others[49,50], the use of participant -observation allowingdirect observation of interaction has advantages oversecond-hand accounts in that what people say in thesocial setting of an interview may be considerably dif-ferent from how they actually behave - and the infor-mation elucidated from mundane events which thesocial actors may not recount to an interviewer [50]can provide valuable insights.Ethics approval for conducting IS-related researchassociated with the cases conducted in British Columbiaand South Africa was granted by the University ofBritish Columbia, with approval in Free State also issuedby the University of Free State. Research in Manitoba hadbeen carried out with ethics approvals from the Univer-sity of Manitobab. Combining formal studies (see foot-notes and references for itemization of various surveysconducted, key informant interviews, and publicationsconsulted) with two decades of participant-observationsfrom members of the research team allows for strongtriangulation.ResultsThe results from the C-M-O analysis of the four casestudies are presented below and summarized in a tabularformat, with a further summary provided according tothe micro, meso and macro-level context of each case:The Experience in Canada1a. Manitoba (1990–1999)A database on occupational health for healthcare wasconstructed in the province of Manitoba, during the1990s – the first of its kind in Canada [51]. As summa-rized in Table 2, rudimentary technology was used in apartnership between our then University of Manitoba-based team and the Winnipeg Health Sciences Centre(HSC), at that time Canada’s largest integrated healthcarecomplex, to collect data related to causes and contribu-tors to work-related disease and injury in healthcareworkers [52], as well as immunization rates [53], and theeffectiveness of interventions [54]. Considerable successwas achieved [55]. The Winnipeg HSC was well resourcedwith a strong occupational health team and a functionaljoint worker-management OHS committee. The occupa-tional health practitioners served as resource-experts exofficio to the bipartite committee, and as such, weretrusted by both workplace parties.Thus, at the micro level (personal relationships on aday-to-day basis), the fact that the database was designedand governed within a clear university-hospital partner-ship headed by a single individual with dual responsibility(co-author AY) made the use of the system both for re-search and operations quite smooth; at the meso level(stakeholder relations within the hospital), interactionswere cooperative, with power imbalances between work-ers and employers adequately mitigated by the OHS Com-mittee (the “agent” in this regard); and at the macro level(general socioeconomic and political conditions), the con-text was very supportive for IS introduction with the pro-vincial social democratic government quite enabling.Table 2 summarizes the key context-mechanism-outcomerelationships that the authorship team synthesized fromthe various articles written, and participant-observerexperiences. There was no overt power struggle that inter-fered with the development, implementation or use of theIS, albeit the tool was limited by lack of technologicalsophistication.1b. British Columbia (2000–2010)In 1999, the Occupational Health and Safety Agency forHealthcare (OHSAH) was created in another CanadianTable 2 Context-mechanism-outcome (C-M-O) summary of workplace health IS use - Manitoba, CanadaContext Surveillance system - created to assist occupational health (OH) department’s health professionals in a large (7,000+ workers)well-resourced teaching hospital with their primary and secondary prevention activities as well as for implementation research(database not containing fields necessary for claims cost containment).Bipartite (union-employer) health and safety (H&S) committee supportive; labour relations amicable.Mechanism Governance: Developed under auspices of a university-hospitalpartnership (which informed a bipartite H&S committee), with anaffiliation agreement in place for collaborative research and service.Technology: Easily accessible; existing standard software (Excel, Access databases).Outcome Decrease in injury rates and time loss due to injuries demonstrated; information also used to improve vaccination programs,and foster good research.Conclusion Useful and sustainable, albeit limited to one workplace as system not web-based, and screens not optimally user-friendly,so required commitment to data collection and data entry.“Tool”, but not a highly efficient one. The power dynamics were such that the risk that the IS would be used as a“weapon” was minimal.Spiegel et al. BMC Medical Informatics and Decision Making 2012, 12:84 Page 5 of 15http://www.biomedcentral.com/1472-6947/12/84province, British Columbia (BC), as a bipartite union-employer collaborative agency, and programs weredeveloped from this perspective [56]. In designing a spe-cific database to monitor and evaluate overhead lifts toprevent musculoskeletal injuries to healthcare workers[55], it became clear that prevention requires more thantracking injuries – it calls for occupational health practi-tioners and active involvement by frontline workers.This lesson, regarding the importance of carefully con-sidering the perspective amongst the different partiesalso came across clearly in the evaluation of the Preven-tion and Early Active Return-to-Work Safety program[57,58] in which an analysis late in the program showedthat the program theory differed considerably amongstthe worker, the employer and the practitioner stake-holders [59]. Building on the experience in Manitobaand advancements in the internet, a web-based systemwas developed, called Workplace Health InformationTracking and Evaluation (WHITE) [60,61]. The context,however, was quite different. Although BC is a wealthierprovince than Manitoba, according to a formal needsassessment of occupational health resources conductedfor a provincial agency, occupational health staffing was“deplorably” lower than international norms [62], labourrelations were more volatile, the power imbalancesmore pronounced and the governance of WHITE wasunclear.Interviews with stakeholders conducted more than fiveyears after implementation of WHITE demonstratedstrong commitment to the integration of occupationalhealth data collection across health regions in BC tocomply with reporting regulations, and to improve com-pensation claim and disability management. However,there was an apparent lack of collaboration between sys-tem users, management, and system developers. More-over, it seemed to us that the employers were moreinterested in the claims management aspect to reducecosts than using the data to promote bipartite collabor-ation supporting workplace hazard reduction [61], letalone empowering the workforce. In 2010, OHSAH wasdiscontinued as a bipartite agency, and WHITE, with noclear rights entrenched for researchers or frontlineworkers to access the data, was transferred to the IT de-partment of the provincial health authority.Thus, at the micro-level, the fact that the database wasdesigned and governed at the provincial level, but thedata were meant to be gathered and used at the (under-staffed) health facility level was problematic; the mesolevel was complicated by unclear governance regardingrights to access and use of data; and at the macro level,the conservative political environment and weak unionsmade it difficult to use the system for empoweringthe workforce. Table 3 summarizes the key context-mechanism-outcome synthesized for this case. Thepower asymmetries in this case did interfere with use ofthe IS; albeit the IS served as a useful tool to controlclaims costs, its usefulness to OHS committees, as notedpreviously [61], was more problematic.Comparing the asymmetries of power in labour rela-tions between the BC and Manitoba experience pointsto the importance of ensuring that the purpose of thedata system be carefully monitored to ensure that its useis not diverted to alternative objectives (e.g. from initiallyempowering workforces to take action to improve theirhealth and safety, toward facilitating employers’ ability toreduce claim costs) [61].Table 3 C-M-O summary - British Columbia, CanadaContext Workplace health information tracking and evaluation system (WHITE) developed within a bipartite healthcare agency forall health sector sites in a wealthy province, but site-specific OH departments poorly resourced.Labour relations volatile and unions relatively weak in this period.With time, collaboration between system users (practitioners), management and system developers became weaker,with no systematic use by H&S committees.Mechanism Governance: Developed by a university researcher working with bipartite provincial agency, but no affiliationagreement between the agency and university partner; later the IS was transferred from the bipartite agency toemployer-control, with no rights to access for researchers.Technology: Customized web-based software developed, using proprietary database technology.Outcome Decrease in injury rates and time loss due to injuries demonstrated; information has not been able to be sustainablyused to improve workforce health, or initiatives such as vaccination programs, nor foster ongoing research.Conclusion Useful to employers for ongoing claims management,Limited use of data either to promote bipartite collaboration for reducing workplace hazards or to support programs thatrequire OH staff in place, or for ongoing high quality research.Web-based system made it useful across entire province, including multiple workplaces, but expensive to maintain.“Tool” for employers and regulators;“Weapon” to busy OH practitioners who are stretched to their limit with no time for data entry, and possibly to workersconcerned that the greater ‘efficiency’ in absenteeism control and time-loss reduction could hurt vulnerable workers in aclimate of weaker job security.Spiegel et al. BMC Medical Informatics and Decision Making 2012, 12:84 Page 6 of 15http://www.biomedcentral.com/1472-6947/12/84Collaboration in South AfricaWhile concerns about the healthcare workforce are glo-bal, the greatest need for health workers and their pro-tection is in Africa, where the WHO estimates that theworkforce will need to increase by 140% before a criticalshortage is overcome [35]. Furthermore, health workersare at higher risk than the general population for tu-berculosis and other infectious diseases [63], wheresuch illness is creating additional staff shortages [64],contributing to increased stress, burn-out and relatedhealth risks [65]. The increased workload is in part at-tributable to the high burden of HIV/AIDS in the coun-try as well as the risk for infection to health workers(a national HIV prevalence of 15. 7% among South Africa’snurses [66]). Meanwhile HICs have actively encouragedemigration of health workers [67], only aggravating a diffi-cult situation. In addition to the fact that South Africacurrently faces a major crisis in terms of human resourcesfor health, there are skewed distributions between thepublic and private sectors [68].Building upon the experience in the two contexts inCanada, and from interactions through the WHO col-laborating centre network [69] where the collaboratingpartners first met, the WHITE system was re-designedand re-named the Occupational Health And Safety In-formation System (OHASIS). Unlike its Canadian prede-cessor, OHASIS explicitly provided modules to helpbuild capacity of joint worker-employer OHS commit-tees and to improve working conditions; and unlike itsCanadian predecessor OHASIS was not built with aclaims management module. The explicit purpose ofOHASIS is to provide workplace and workforce surveil-lance data for operational decision-making in improvingworking conditions and over-all worker health, in linewith government policy in this regard, while at the sametime serving as a treasure of data for local and inter-national research on determinants of injuries and illnessin health workers.1a. Free State (2006–2011)The pilot launch of the South African OHASIS collabo-ration occurred in one large regional Hospital in Bloem-fontein, Free State, and included input from nationaland provincial decision-makers as well as occupationalhealth practitioners and health worker trade unions [70].This site was selected for several reasons: In 2006, theFree State Department of Health had established a taskteam to coordinate and evaluate occupational health ac-tivities in the province, thereby confirming commitmentfor improving this area; the province had experiencewith information systems in broader population healthapplications [71] and there was strong managementand union support at the local level, as well as interestby the local university. Most importantly, though, thecollaboration between Canadian and South Africanexperts began precisely with a focus on how to build cap-acity of health workers to address their working condi-tions [70] with the data system to target this objective.A feasibility studyc conducted prior to launching a for-mal pilot study evaluation [70] found that severalimprovements were needed in the system– to make itmore user-friendly and robust for data analysis. Thesechanges were then made, but as time had passed, an-other baseline assessment had to be conducted to serveas a comparison when implementing the newly rede-signed system. This second survey, conducted in 2010,assessed frontline workers’ knowledge, practices, andattitudes around infection control and working condi-tions, and was completed by 110 participants at the tar-geted Hospital. In 2012, the survey was re-designed andre-administered to health workers at this hospital andthe sampling frame was expanded to include respon-dents from two other large hospitals located in the FreeState province. The full results with South African colla-borators will be reported elsewhere, but for purposes ofthis analysis it is noteworthy that respondents of the sec-ond survey were divided in their perception of how easyit was to obtain information about potential workplacerisks – 37% found it easy, 39% stated it was not easy,and 22% answered that they did not know. With regardto the dynamics of power within the hospital, contraryto the international team’s impressions that strong sup-port from management was present (possibly in com-parison to the situation in Canada), 41% of 109 staffrespondents reported that they did not feel adequatelysupported by their managers in matters pertaining tohealth and safety; 12% reported that the management/hospital could never be trusted, while 31% said that theycould be trusted a quarter to half of the time, with only38% saying that the management and hospital could al-ways be trusted. This is especially relevant when itcomes to workforce health information; although OHA-SIS has solid security features to guard confidentiality ofpersonal, it is understandable that lack of trust wouldundermine support for the system. Furthermore, manykey informant interviews uncovered themes around theimbalance of power between management and workers.In discussions regarding the divide between workers andmanagers, some workers indicated that while there isorganizational rhetoric around keeping employees safe,in actuality, prevention measures were lacking. Addition-ally, by the time the re-designed system was ready forusage, the political context was no longer optimal forboth implementation.In August 2010, a bitter public sector strike occurred inSouth Africa that dramatically affected the labour situa-tion in the hospitals. Amidst this heightened union-management acrimony, one of the unions decided thatSpiegel et al. BMC Medical Informatics and Decision Making 2012, 12:84 Page 7 of 15http://www.biomedcentral.com/1472-6947/12/84until a properly elected OHS committee was establishedand adequately trained from their perspective, the projectshould not go ahead. Albeit reluctantly, the formal launchof the IS planned for Free State had to be delayed. InSeptember 2011, the issues were resolved, allowing imple-mentation of the new version of OHASIS to go forward.This proceeded well, with logistical deployment issues(e.g. ensuring adequate connectivity among users) thenaddressed. However, in implementing at a second hospitalsite, new challenges emerged, attributable at least in part,to the legacy of racial tensions, as well as possible con-cerns about sharing data from the hospital with the pro-vincial level, where OHASIS was hosted.Thus, at the micro-level, the on-site OH practitionersand the researchers all wanted the IS, there were excellentrelations established and the system was designed forworkforce empowerment. At the meso level, however,considerable barriers occurred due to labour concernsabout governance, which also had to be resolved, andother issues related to power imbalances had to beaddressed through additional clauses to the initial memo-randum of agreement. At the macro level, the dependenceon the Northern partner for sustainable IS technical sup-port emerged as a concern, with the original vision of anongoing relationship with an operating IS system in anHIC setting (i.e. the WHITE system in BC) no longer vi-able. The solution for this came from involving a WHOpartner at the South African national level, as is discussedbelow. Table 4 summarizes the key context-mechanism-outcome relationships the authorship team synthesizedfrom the articles written, interviews, the survey notedabove and the extensive participant-observer experiences.2b. The National Health Laboratory System (NHLS)(2009–2012)With unions at the national level keen to see OHASISadopted, the NHLS did not experience the same politicalconstraints as were experienced in Free State. Accord-ingly, implementation of OHASIS across all 350 labora-tories across South Africa was initiated in mid-2011,operated by the NHLS, with NIOH serving as technicaladvisors. NIOH also agreed to provide technical supportto Free State, which helped resolve both technical andpolitical concerns that arose with the system maintainedfrom Canada, alluded to above. Details of this expe-rience will be presented separately regarding both theexperience within NHLS [72] as well as the implicationsfor the Free State implementation. Table 5 summarizesthe C-M-O analysis conducted from the experience todate.The context of implementing the workforce health in-formation system was thus very different in South Africacompared to Canada. With militant trade unions inhealthcare, attention to union demands had to be a pre-requisite to successful implementation. Moreover, des-pite the fact that the Southern practitioners had initiatedcollaboration with the Northern researchers, the North–South collaboration was still somewhat perceived asthreatening in the less-developed less-well-resourcedsetting, but was openly welcomed in the better-resourced WHO-affiliated centre, NIOH. The mecha-nism for successful implementation was indeed theWHO collaborating centre network, which allowed thebuilding of trust – key for success, and facilitating theempowering of a locally-based WHO collaboratingcentre looking after the IS needs in this field within notonly South Africa but the entire African regiond.Table 6 summarizes our analysis of how factors relatedto context and mechanisms affected the contribution ofIS to achieving outcomes in the four cases presented.The specific characteristics, outlined in the form of achecklist in Table 7, suggest criteria that should be con-sidered when implementing technical solutions to ad-dress social health determinants generally.Table 4 C-M-O summary - Free State, South AfricaContext IS (OHASIS) developed to increase capacity for improving working conditions in South Africa healthcare, launched as pilot inresource poor setting but with more human resources devoted to OH than in BC (comparable to the Manitoba setting).Initially strong bipartite H&S committee support, but political changes, heightened racial tensions, increased unionmilitancy and complex governance concerns created challengesMechanism Governance: Partnership between Canadian and South African university-based researchers, with provincial health department.Technology: Similar to the BC system, but with module development emphasizing prevention and capacity-building OH activities.Outcome While feasibility study was positive, and OH professionals keen to use system, implementation of revised system was delayeddue to political power struggles (union discontent with how joint health and safety committees were established).Conclusion System designed for prevention and empowerment of the workforce; delay in implementation because of expressed union concerns.“Tool” to OH practitioners, however efficiencies not realized, as new system with improved reporting functions was neverimplemented in this time frame.“Weapon” use by union militants to leverage achieving other demands.“White elephant” to the researchers and decision-makers who invested in the system, and so far do not have a usablesystem implemented.Spiegel et al. BMC Medical Informatics and Decision Making 2012, 12:84 Page 8 of 15http://www.biomedcentral.com/1472-6947/12/84DiscussionBuilding on the observation that IS can change theorganizational landscape of power and status [73], Tjora& Scambler [9] suggest that a factor explaining why ISinnovations have been disappointing lies in IS havingignored the importance of “role” within an organization.Our analysis of factors affecting IS implementation, con-sistent with that of Sein and colleagues [74] also showsthat the distinct interests and roles of different sets ofactors are critical in shaping IS introduction and use –and affect the dialectic that our adapted structurationistframing of IS implementation seeks to explain.Amid mounting interest not only in the health of healthworkers but more broadly in the significance of socialdeterminants of health, systematic consideration of howdata systems affect power relations is warranted. Innova-tions such as IS applications for conducting inspection ac-tivities [75] providing more comprehensive and timelyinformation to health professionals and managers [76] orestablishing more comprehensive general surveillance ofaffected communities [77] suggest ways that experts canextend their consideration of health-relevant information.Nonetheless, direct engagement of the people whose healthis being monitored has remained neglected – leaving themde facto as little more than objects for observation.There is worldwide consensus on the need to improvemonitoring and evaluation of the health and safety ofhealth workers, especially given the unrelenting chal-lenges of HIV and tuberculosis [78]. As for health infor-mation systems generally, leadership and drive hasprimarily come from professional, expert or managerialchampions, often acting in concert with proprietarycommercial developers and providers [79]. A review ofcommercial software systems implemented to addressthe health sector workforce, in fact observed that thesehave tended to focus on providing information for finan-cial and administrative managers, with limited usefulnessto surveillance of workplace risks [80]. Our analysis ofthe power relations helps explain the skewed develop-ment of IS in this setting, and the importance of newinitiatives underway internationally.The incentive structures reinforced by how intellectualproperty and issues related to licensing have defined thecontext for developing IS solutions for OHS issues mosthighly valued by financial and administrative managersmust, however, also explicitly be taken into account. ToTable 5 C-M-O summary - National Health Laboratory System, South AfricaContext Same tool (OHASIS) adapted to increase capacity in healthcare laboratories across South Africa to improve working conditions.Personnel accustomed to computerized data collection, resources devoted to OH staffing, and union support present atthe national level.Mechanism Governance: Partnership between Canadian and South African researchers affirmed in writing.Technology: Similar to above two, but with more user-friendly features.Outcome Being implemented in 350 laboratories across South Africa, but sustainability of system still questionable, as IT department stilldepends on northern partner, but transition plan in place.Conclusion Even when a system is successfully launched, and labour relations are supportive, IT capacity-building is essential from theoutset to ensure sustainability.“Tool” to all, but risks becoming a “White elephant” if the IT capacity can not be quickly built to take over full maintenance andfurther development.Table 6 Summary of contexts and mechanisms needed for successful outcome at different levelsMan BC FS NHLSMicroCommitment to health and safety including adequatestaffing to plan, implement and evaluate interventions+ - + +MesoClear governance (access, use of data) + - + +Good labour relations so neither side is motivated to usethe IS as a “weapon” rather than “tool”+ - - +MacroEnabling political environment + - + +Sustainable local IT capacity (so system does notbecome white elephant)- + -* +*based on the analysis conducted for this study, efforts are now underway to implement mechanisms for sustaining needed capacity, in the form of internationalnetworks and regional WHO Collaborating Centres taking on leadership roles in their jurisdictions.Spiegel et al. BMC Medical Informatics and Decision Making 2012, 12:84 Page 9 of 15http://www.biomedcentral.com/1472-6947/12/84ensure that the potential benefits of IS solutions are notlimited or skewed by proprietary ownership, IS innova-tions have also triggered development of alternative li-censing and knowledge-sharing orientations. As noted[69], in the spirit of seeking such solutions with inter-national partners and mitigating potential power imba-lances, our team has explicitly pursued the model ofCreative Commons licensing described above, and is ac-tively pursuing further software refinements with theSouth African and other International partners.Exploring further implementation opportunities, in2009, our University of British Columbia-based teamjoined an international collaboration linking the PanAmerican Health Organization (PAHO), Health Canada,the Ecuadorian Ministry of Public Health, and VancouverCoastal Health Authority (VCH) to strengthen Ecuador’scapacity to promote healthier and safer healthcare work-ing conditions [81]. The project fit under the frameworkof the PAHO Regional Plan on Workers’ Health [82], aswell as the WHO Global Plan of Action on Workers’Health [83], which explicitly urges member countries toimplement occupational health and safety policies andprograms in the healthcare sector. The interdisciplinaryteam selected three public Ecuadorian hospitals and con-ducted a baseline assessment in thirteen medical unitsfrom across these three facilities [84]. Building on a well-established collaboration between the Canadians andEcuadorians [85] efforts were pursued with governmentand hospital personnel and academic partners to adaptthe IS for health workers in Ecuador, linking this withTable 7 Power-relations checklist for implementing occupational health information systems (IS)MICRO: WITHIN THE WORKPLACE OHS DEPARTMENT –WHO MIGHT BENEFIT OR FEEL THREATENED BY THE IS?□Will only specific occupational health practitioners enter data and access system? Or all? If only some, is there good consensus on this amongst theOHS personnel?□What will be the impact on staff workload? Is staffing adequate?□Are personnel adequately trained to capture data correctly?□Will health and safety representatives be able to access any aspects of the system?□ If so, are they trained adequately?□Are all appropriate personnel trained to interpret and act on the data?□Who will receive aggregated reports?□How often will aggregated reports be generated? Who will write commentaries?□ Is the local technology adequate – (i. e. computers, bandwidth, etc. )?□Have policies and procedure been written to guide system use, confidentiality of data, and access to reports?□ Is there a communications plan established between system implementers and the workplace staff who will use the system?MESO: WITHIN THE ORGANIZATIONWHO MIGHT BENEFIT OR FEEL THREATENED BY TH IS?□Have the unions or worker representatives been adequately consulted about the introduction of such a system?□Were frontline managers adequately consulted about the introduction of such a system?□Was the information technology department of the workplace adequately involved?□Did all the appropriate workplace parties have input to the design, policies and procedures regarding use of the system?□Are all the workplace parties throughout the organization aware of how they might benefit from the system?□Are there clear channels of communication between units within the organization to ensure equity and foster shared involvement/ownership?MACRO: BEYOND THE ORGANIZATIONWHO MIGHT BENEFIT OR FEEL THREATENED BY THE IS?□Who designed the system? If the design occurred out of the jurisdiction where the IS is being implemented, were local stakeholders adequatelyinvolved in adaptations?□ Is the governance of system use and financing clear?□Who governs the maintenance, upgrade, or system design modifications?□How is the maintenance and upgrade of the IS being financed? Are the financial benefits fair?□Are the terms and conditions sustainable even if the current decision-makers and/or technical personnel all change?□What aspects of the local, regional, national or international political climate may impact the system? If a less worker-friendly government come in,will this impact system use?□ If financial issues and governance involve multi-scalar (i. e. hospital- province/state –national-international) cooperation, what other political issuesmay arise and how can these be managed?Spiegel et al. BMC Medical Informatics and Decision Making 2012, 12:84 Page 10 of 15http://www.biomedcentral.com/1472-6947/12/84support provided by the Andean Region Health Observa-tory maintained by the Andina Simon Bolivar University[86]. Other countries have expressed interest in OHASIS,and efforts are now underway to make this availablethrough our Ecuadorian-based partner, using a CreativeCommons license encompassing the three principlesnoted above.Meanwhile, the US National Institute of OccupationalSafety and Health (NIOSH) very recently created a por-tal to collect information in a central repository – notonly from within the US, but internationally; OHASISwas then re-designed to be able to not only export intothe NIOSH repository, but the OHASIS team has cre-ated a user-friendly interface to facilitate the adaptationof other systems to contribute to this data pool to facili-tate international collaboration. It is also noteworthy thatOHASIS contains EpiNettm within it [87] – modulesdeveloped to specifically capture information on needlestick injuries and other blood and body fluid exposures –used freely in dozens of countries worldwide, also opera-ting within the WHO network.The findings from our realist analysis of two decades ofexperience show the extreme relevance of consideringpower relations in the micro, meso and macro contexts ofhealthcare workplace setting, where power dynamics areof considerable importance, and access to information onworkplace hazards, occupational injuries and illnesses, aswell as information about the risk factors, and health ofhealth workers can be quite contentious [61]. The lessonslearned in this regard are summarized in Table 7 in theform of a checklist for consideration in pursuing imple-mentation. Although experts have previously noted thatignoring political realities and organizational social dy-namics can lead to IS failure [88], this is the first studythat examined these factors in determining IS use to pro-mote workforce health.Efforts to make information about occupational healthrisks and their controls available online are increasinglybeing embraced [37], in keeping with evidence that em-powerment of the workforce (and OHS committees) isof paramount importance in improving workplace con-ditions and worker well-being [89], and requires properinformation and training [90]. Furthermore, researchshows an important link between trust, managementsupportiveness and safety culture [91]. Accordingly, webegan our work in Free State with workplace audits, andassessments of worker knowledge, attitudes and self-reported practice, and then incorporated a module onworkplace assessment into the IS itself, thereby addres-sing a specific surveillance need noted by Gaydos andcolleagues [75]. Our analysis, nevertheless, suggests thateven attempts to address this weakness can still be insuf-ficient to overcome the impact of the power asymme-tries in workplace settings.ConclusionAs we argued more than twenty years ago [92], a surveil-lance system is more than a data collection system – itrequires the ability to interpret and act on the informa-tion as part of an inherently iterative monitoring func-tion. This capacity-strengthening focus is what makes anIS in this setting such a potentially powerful tool.Nevertheless, meticulous attention is needed to ad-dress power asymmetries at different levels: micro (e.g.worker – management relations within the workplace,including addressing the disincentives to reportinghazards and incidents, empowerment of OHS commit-tees and adequate staffing); meso (including governanceissues, role of university partners and North–south teamdynamics in the absence of sufficient capacity-building);and macro (including political climate, proprietary li-censing and global patent protections). North–South-South partnerships (i.e. where an HIC partner facilitatesdirect collaborations between LMIC partners) should beencouraged, and are showing promise. In this regard,making IS solutions that are developed in HIC settingsavailable for application and further collaborative devel-opment in LMIC contexts, as was initially planned, cansubstantially reduce cost burdens and broaden a “cre-ative commons” network and community of practice toenhance sustainability of IS uses in LMIC settings.If the power asymmetries in the mechanisms estab-lished for implementation are not properly addressed,however, the system, while meant as a tool, can be seenas a weapon (an instrument of control exerting power ofone set of agents over another). Such perception cancome from many sources: by busy healthcare profes-sionals who feel pressured into collecting and enteringdata into a system that provides them minimal benefit;by a workforce that is skeptical that confidentiality willbe maintained and/or concerned that the informationwill be used to empower managers as absenteeism po-lice; or indeed, by union activists, who may wish to usethis tool for purposes which the system was not meantto address. Furthermore, of particular importance inthese resource-constrained times, the system, with itstechnologically sophisticated requirements in settingsnot habituated to this, can easily become a white ele-phant. A structurationist analysis, reinforced by explicitconsideration of these power dynamics, applied at micro,meso and macro levels, is therefore especially usefulwhen embarking on IS implementation.EndnotesaWe adopt the World Health Organization definition:“Health workers are all people whose main activities areaimed at enhancing health. They include the peoplewho provide health services – such as doctors, nurses,pharmacists, laboratory technicians – and managementSpiegel et al. BMC Medical Informatics and Decision Making 2012, 12:84 Page 11 of 15http://www.biomedcentral.com/1472-6947/12/84and support workers such as financial officers, cooks,drivers and cleaners. "WHO Fact Sheet #302. April2006 (http://whqlibdoc. who. int/fact_sheet/2006/FS_302.pdf). Nonetheless, the shortage of health workers globallydoes not apply to all categories of health workers.b This includes UBC ethics certificate # H10-00360-A003 and H05-80551. Other certificate numbers, includ-ing those for studies more than 5 years old, are availableon request.c In September 2007, the NIOH and the South AfricanHealth Department organized a seminar attended by na-tional and provincial occupational health coordinatorsand Canadian team members. Almost 100 participantsattended this initial workshop, including health andsafety representatives, occupational health and infectioncontrol professionals, union activists, clinicians, man-agers and academic collaborators. Here, a standard corecurriculum, core competencies, and a methodology forrolling out a training program for health and safetyrepresentatives in South Africa was discussed. The re-search and training materials were developed based onexpertise gained from work conducted by team membersin Canada, South Africa and Ecuador and included atraining guide, problem-based learning case scenariosand a workplace audit tool. Topics covered in the train-ing included: roles and responsibilities of occupationalhealth professionals and health and safety committees,basic concepts in occupational health (i.e. the hierarchyof control measures), incident reporting and investiga-tion, workplace assessment and workforce health. Thesematerials were subsequently used each year to train newhealth and safety committee members at Pelonomi Hos-pital. All workshops were formally evaluated and materi-als and format were amended as necessary.d A memorandum of agreement was signed November30, 2011 between the University of British Columbia andthe South African NHLS to allow free transfer of thesystem, including source code, with the explicit under-standing that the South African partner would providesupport for OHASIS implementation and use across theAfrican region.Appendix Data Sources Details (notes for Table 1)a These studies, ranging from analyses of the epidemi-ology of injuries to various occupational groups of healthworkers (nurses, food handlers, laundry workers, etc.) todeterminants of various types of injuries (needlesticks,musculoskeletal injuries, violence, etc.) to effectivenessof various programs (return-to-work programs, muscu-loskeletal injury prevention initiatives, vaccine programs,etc.) were reviewed by at least two members of the re-search team. In addition, articles that describe the con-text were reviewed, including those that detail thehistory and challenges in the Canadian Institutes forHealth Research (CIHR) Community Alliances forHealth Research (CAHR) program that funded much ofthe collaborative research with the institutions involvede. g. [56,93] were also reviewed.b These included surveys prior to implementation(e. g. of managers at Fraser Health in BC [94], and offront-line health worker e. g. [95]). In addition to thesurveys referenced in the text related to the NHLS im-plementation, a separate publication will be submittednext year detailing the results of the survey conducted atthree hospitals in Free State.c Key informants specifically interviewed in formalinterviews included senior managers as well as personnelin charge of prevention in BC; in South Africa, occupa-tional health practitioners were formally interviewed aswell. Specifically, co-author CD conducted one-hoursemi-structured key informant interviews in August2010 with two occupational health champions workingin Bloemfontein, South Africa. Due to the nature of theinformation solicited, identifying details about the titlesof the informants cannot be given, however, each personworked at two key hospitals in the area. One personworked in OSH and was assisting in the implementationof OHASIS at their hospital, while the other providedcare to health workers who experienced OHS events inthe hospital. Both were able to speak to the complexpower dynamics that both facilitated and impeded theflow of OHS knowledge between health workers, theirunions, management and the executive leaders of theirhospital. While our team led research and implementa-tion planning meetings and activities, two project teammembers observed interactions between staff of differentlevels over a full month in August 2010. This qualitativedata was also thematically analyzed. Co-author AW con-ducted one-hour semi-structured key informant inter-views with various professionals involved in theimplementation and use of the IS in BC in April andMay 2010: interviewees were from two different healthregions and Work Safe BC. (To protect the identities ofthe interviewees, titles and positions in their organiza-tions are not provided here.) AW also conducted a the-matic analysis of the qualitative data. Additionally, AWspent a few months in Free State in 2010 assisting in thedesign of the initial survey and data collection andanalysis.d Co-author AY was the Founding Director of the oc-cupational health department in which the Manitobasystem was developed and implemented (1986–1999), aswell as research director for studies conducted duringthis period using the IS. AY was also the FoundingExecutive Director of OHSAH as well as the PrincipalInvestigator (PI) of the research program in BC (1999–2007) in which the BC IS was developed and implemen-ted. She also is PI of the research program in which theSpiegel et al. BMC Medical Informatics and Decision Making 2012, 12:84 Page 12 of 15http://www.biomedcentral.com/1472-6947/12/84South Africa system was developed and is being imple-mented (2007-ongoing). Her experience and reflectionsconstitute an important part of the fieldwork for thisstudy. Co-author JS is a program evaluation and policyresearcher with over 30 years of experience in variousuniversity and government regulatory and policy (inclu-ding Manitoba and BC) settings in examining effective-ness of occupational and environmental health systemsand interventions. He is PI of a CIHR funded compre-hensive study of the implementation of the informationsystem in South Africa. Co-author KL served as Re-search Manager of the CAHR in BC from 2003–2007,and is now the Research Manager for the research pro-gram in South Africa. Her experience in implementingthe IS and using data collected is thus also an importantsource of insight for this article. Co-author LO con-ducted 10 fieldtrips to South Africa in which the designand implementation of the IS were key topics of obser-vation. She has spent over 9 months total working withoccupational health practitioners, clinical managers, pro-vincial decision-makers and other stakeholders in thisregard.AbbreviationsBC: British Columbia; C–M–O: Context – Method - Outcome; HICs: HighIncome Countries; HSC: Health Sciences Centre (Winnipeg Manitoba Canada);ICOH: International Commission on Occupational Health; IS: InformationSystem; LMICs: Low and Middle Income Countries; NHLS: National HealthLaboratory System (South Africa); NIOH: National Institute of OccupationalHealth (South Africa); OHASIS: Occupational Health And Safety InformationSystem; OHSAH: Occupational Health and Safety Agency for Healthcare (BCCanada); OSS: Open Source Software; PAHO: Pan American HealthOrganization; WHITE: Workplace Health Information Tracking and Evaluation;WHO: World Health Organization.Competing interestsThe authors declare that they have no financial competing interests. Theinformation systems reported herein are non-commercial, and the sourcecode for the Occupational Health And Safety Information System (OHASIS) isfreely available.Authors' contributionsJS is the principal investigator of the “Tool, Weapon or White Elephant” studyfunded by the Canadian Institutes of Health Research, which helped formthe conceptual basis of this article. He led in the conceptualization andwriting of this manuscript, along with AY. He also personally did the finalediting and preparation for submission. KL conducted a literature review tosupport the article, assisted in the initial drafting and contributed herobservations from years of working with the information system in BritishColumbia. LO assisted with the drafting of the article based on herinvolvement with the system development and implementation in SouthAfrica. AW assisted with the drafting of the article, as well as drafting andconducting the survey of system users in South Africa noted in this article,along with the collation and interpretation of the results; she alsointerviewed system users in British Columbia and incorporated theseobservations into the article. CD also assisted with the drafting of the article,as well as drafting and conducting the survey of system users in SouthAfrica. She also conducted interviews of system users in South Africa andcontributed these observations to the article. AY, the senior author on thismanuscript, played the lead role in developing, adapting and implementingthe systems described in all five settings, as well as conducting researchrelated to system use in all five settings. She wrote the first draft of thisarticle, and worked closely with Dr. Spiegel on the final draft. All authorsread and approved the final manuscript.AcknowledgementsWe are grateful for the financial support provided by the Canadian Institutesof Health Research (CIHR). We thank the faculty and staff at the University ofBritish Columbia, for their assistance including Justin LoChang, StephenBarker, and Joe Tremblay, as well as colleagues, and the health workers anddecision-makers in all five of the settings described who shared their viewsand allowed us to capture their experience.Received: 14 November 2011 Accepted: 28 June 2012Published: 6 August 2012References1. Ammenwerth E, de Keizer N: An inventory of evaluation studies ofinformation technology in health care: trends in evaluation research1982–2002. Methods Inf Med 2005, 44:44–56.2. Goldzweig CL, Towfigh A, Maglione M, Shekelle PG: Costs and benefits ofhealth information technology: new trends from the literature. Health Aff(Millwood) 2009, 28(2):w282–w293.3. Heeks R: Health information systems: Failure, success and improvisation.Int J Medical Informatics 2006, 75(2):125–137.4. Littlejohns P, Wyatt JC, Garvican L: Evaluating computerised healthinformation systems: hard lessons still to be learnt. BMJ 2003,326(7394):860–863.5. Avgerou C: The significance of context in information systems andorganizational change. Inf Syst J 2001, 11:43–63.6. Rahimi B, Vimarlund V: Methods to evaluate health information systemsin healthcare settings: a literature review. J Med Syst 2007, 31(5):397–432.7. Giddens A: The constitution of society: Outline of the theory of structuration.Cambridge: Polity Press; 1984.8. Orlikowski WJ: The duality of technology: Rethinking the concept oftechnology in organizations. Organisational Sci 1992, 3(2):398–427.9. Tjora AH, Scambler G: Square pegs in round holes: Information systems,hospitals and the significance of contextual awareness. SocSci Med 2009,68(3):519–525.10. Jones MR, Karsten H: Giddens’s Structuration Theory and InformationSystems Research. MIS Quarterly 2008, 32(1):127–157.11. Myers M, Klein H: A set of principles for conducting critical research ininformation systems. MIS Quarterly 2011, 35:17–36.12. Ghose R: Community Participation, Spatial Knowledge Production, andGIS Use in Inner-City Revitalization. J Urban Technol 2003, 10(1):39–60.13. Sieber R: GIS Implementation in the Grassroots. J Urban and RegionalInformation Systems Assoc 2000, 12(1):15–51.14. Doolin B: Power and resistance in the implementation of a medicalmanagement information system. Inf Syst J 2004, 14:343–362.15. Martin CM: Understanding and changing health systems–an instinctiveand natural process? J Eval Clin Pract 2009, 15(5):859–860.16. O’Mahony S: Guarding the commons: how community managedsoftware projects protect their work. Research Policy 2003,32(7):1179–1198.17. Hertel G, Niedner S, Herrmann S: Motivation of software developers inOpen Source projects: an Internet-based survey of contributors to theLinux kernel. Research Policy 2003, 32(7):1159–1177.18. Kahen G: Assessment of Information Technology for DevelopingCountries: Appropriateness, Local Constraints, IT Characteristics andImpacts. Int J Comput Applications in Technologies 1995, 5(5/6):325–333.19. Kimaro H: Strategies for Developing Human Resource Capacity toSupport Sustainability of ICT Based Health Information Systems: A CaseStudy from Tanzania. The Electronic J Information Systems in DevelopingCountries 2006, 26(2):1–23.20. Avgerou C: Information systems in developing countries: a criticalresearch review. J Inf Technol 2008, 23:133–146.21. Heeks R: Information Systems and Developing Countries: Failure, successand local improvisations. Inf Soc 2002, 18(2):101–112.22. Chandrasekhar CP, Ghosh J: Information and communication technologiesand health in low income countries: the potential and the constraints.Bull World Health Organization 2001, 79(2):850–855.23. Walsham G, Sahay S: Research on information systems in developingcountries: current landscape and future prospects. Inf Technol Dev 2005,12(1):7–24.24. Guba E, Lincoln Y: Fourth generation evaluation. Newbury Park, California:Sage; 1989.Spiegel et al. BMC Medical Informatics and Decision Making 2012, 12:84 Page 13 of 15http://www.biomedcentral.com/1472-6947/12/8425. Boonstra A, Govers M: Understanding ERP system implementation in ahospital by analysing stakeholders. New Technology, Work andEmployment 2009, 24:177–193.26. Melling J, Booth A: Managing the Modern Workplace: Productivity. Politicsand Workplace Culture in Postwar Britain: Ashgate Publishing, Ltd.; 2008.27. O'Grady J: Joint health and safety committees: Finding a balance. In Injuryand the new world of work. Edited by Sullivan T. Vancouver, BC: UBC; 2000.28. Bryce G, Manga F: The effectiveness of joint health and safetycommittees. Relations Industrielles 1985, 40(2):257–282.29. Boden L, Hall J, Leverstein C, Punnett L, Boden L, Hall J, Leverstein C,Punnett L: The Impact of Health and Safety Committees: A Study Basedon Survey, Interview, and Occupational Safety and Health AdministrationData. J Occup Environ Med 1984, 26(11):829–834.30. Reilly B, Paci P, Holi P: Unions, safety committees, and workplace injuries.Br J Ind Relat 1995, 33(2):275–289.31. Pragnell B: Occupational & Health Committees in NSW: An Analysis for theAWIRS Data. In. Sydney: Australian Centre for Industrial Relations Researchand Teaching; 1994.32. Frick K, Walters D: Worker representation on health and safety in smallenterprises: Lessons from a Swedish approach. Int Labour Rev 1998,137(3):367–389.33. Gallagher C, Underhill E: Managing work health and safety: recentdevelopments and future directions. Asia Pacific J Human Resources 2012,50(2):227–244.34. Myers J, Macun I: The sociologic context of occupational health in SouthAfrica. Am J Public Health 1989, 79(2):216–224.35. World Health Organisation (WHO): The world health report, 'Chapter 4 -Selected occupational risks'. 2004, http://www.who.int/whr/2002/chapter4/en/index8.html.36. Hongoro C, McPake B: How to bridge the gap in human resources forhealth. Lancet 2004, 364:1451–1456.37. Rhebergen MD, Hulshof CT, Lenderink AF, van Dijk FJ: An online networktool for quality information to answer questions about occupationalsafety and health: usability and applicability. BMC Med Inform DecisMak2010, 10:63.38. Milgate N, Innes E, O'Loughlin K: Examining the effectiveness of healthand safety committees and representatives: a review. Work (Reading,Mass 2002 2002, 19(3):281–290.39. Spiegel JM, Labonte R, Ostry AS: Understanding "globalization" as adeterminant of health determinants: a critical perspective. Int J OccupEnviron Health 2004, 10:360–367. Special Edition Spring.40. Gershon RR, Sherman M, Mitchell C, Vlahov D, Erwin MJ, Lears MK, Felknor S,Lubelczyk RA, Alter MJ: Prevalence and risk factors for bloodborne exposureand infection in correctional healthcare workers. Infect Control HospEpidemiol2007, 28(1):24–30.41. Schmid K, Schwager C, Drexler H: Needlestick injuries and otheroccupational exposures to body fluids amongst employees and medicalstudents of a German university: incidence and follow-up. 2007,65(2):124–130.42. Bryce E, Yassi A, Nophale L, Scharf S, Dybka L, Lavoie M-C, Tennassee M,Fujii R: International adaptation and sharing of an infection control andoccupational health inspection tool for healthcare facilities, Bringing lessonsback home. Cape Town, South Africa: International Commission onOccupational Health Congress; 2009.43. Yassi A, Nophale L, Dybka L, Bryce E, Kruger W, Spiegel J: Building capacityto secure healthier and safer working conditions for healthcare workers:A South African-Canadian collaboration. Int J Occup & Environ Hygiene2009, 15(4):369–380.44. O'Hara L: Occupational Health Surveillance in Health Care: A follow-up to the Morocco workshop. In ICOH- Occupational Health for HealthCare Workers Scientific Committee Virtual Seminar. Edited by. 2011.Available from: http://innovation.ghrp.ubc.ca/icoh/ICOH_surveillance%20workshop_FINAL3.pptx45. Pawson R, Tilley N: Realistic Evaluation. London, UK: Sage; 1997.46. Pawson R, Greenhalgh T, Harvey G, et al: Realist review: A new method ofsystematic review designed for complex policy interventions. J HealthServices Res 2005, 10(Supplement 1):21–34.47. Astbury B, Leeuw FL: Unpacking Black Boxes: Mechanisms and TheoryBuilding in Evaluation. Am J Evaluation 2010, 31:363.48. Porter S: A participant observation study of power relations betweennurses and doctors in a general hospital. J Adv Nurs 1991, 16(6):728–735.49. Kawulich BB: Participant Observation as a Data Collection Method. Forum:Qualitative Social Res 2005, 6(2):43.50. Demunck VC, Sobo EJ: Using methods in the field: a practical introductionand casebook. Walnut Creek, California: AltaMira Press; 1998.51. Yassi A, Tate R, Cooper J, Jenkins J: Causes of staff abuse in a large innercity hospital: Implications for prevention. In The International Congress ofOccupational Health. Edinburgh, Scotland: Occupational Health for HealthCare Workers Conference; 1997.52. Yassi A: Utilizing data systems to develop and monitor occupationalhealth programs in a large Canadian hospital. Methods of Information inMed 1998, 37:125–129.53. Yassi A, Khokhar J, Marceniuk M, McGill M: Hepatitis B vaccination forhealth care workers: Evaluation of acceptance rate and programstrategy at a large tertiary care hospital. Canadian J Infection Control1993, 8:94–97.54. Cooper J, Tate R, Yassi A: Work hardening in an early return to workprogram for nurses with back injury. Work 1997, 8:149–156.55. Spiegel JM, Yassi A, Ronald LA, Tate RB, Hacking P, Colby T: Implementing aresident lifting system in an extended care hospital. Demonstratingcost-benefit. Am Assoc Occup Health Nurses J 2002, 50(3):128–134.56. Yassi A, Ostry AS, Spiegel J, Walsh G, de Boer HM: A collaborativeevidence-based approach to making healthcare a healthier place towork. Hospital Quarterly 2002, 5(3):70–78.57. Badii M, Keen D, Yu S, Yassi A: Evaluation of a comprehensive integratedworkplace-based program to reduce occupational musculoskeletal injuryand its associated morbidity in a large hospital. J Occup Environ Med2006, 48(11):1159–1165.58. Davis P, Badii M, Yassi A: Preventing disability from occupationalmusculoskeletal injuries in an urban, acute and tertiary care hospital:results from a prevention and early active return-to-work safelyprogram. J Occup Environ Med 2004, 46(12):1253–1262.59. Maiwald K, de Rijk A, Guzman J, Schonstein E, Yassi A: Evaluation of aWorkplace Disability Prevention Intervention in Canada: ExaminingDiffering Perceptions of Stakeholders. J Occup Rehabilitation 2010,179–89:1573–3688.60. Gilligan T, Alamgir H: Bridging the Gap: An innovative surveillance systemto monitor the health of British Columbia’s healthcare workforce.Canadian J Publ Health 2008, 99(Suppl 2):478–482.61. Spiegel J, Lockhart K, LoChang J, Tremblay J, Dybka L, Yassi A: Workplaceand workforce health information systems in healthcare: Acknowledgingthe role of university researchers and highlighting the importance ofhealth and safety committee capacity-building. Canadian J Publ Health2009, 100(2):157.62. Pollock S, Yassi A, Connell I, Gamage B, Copes R: Are you prepared?Defining occupational health resource needs to prevent infectiousdisease transmission in the health care sector. Healthcare Manage Forum2009, 22(1):52–56.63. Menzies D, Joshi R, Pai M: Risk of tuberculosis infection and diseaseassociated with work in health care settings. The Int J Tuberculosis andLung Disease 2007, 11(6):593–605.64. Daviaud E, Chopra M: How much is not enough? Human resourcesrequirements for primary health care: a case study from South Africa.Bulletin of the World Health Organization 2008, 86(1):46–51.65. Gounden Y, Moodley J: Exposure to human immunodeficiency virusamong healthcare workers in South Africa. Int J Gynaecol Obstet 2000,69(3):265–270.66. Shisana O, Hall EJ, Maluleke R, Chauveau J, Schwabe C: HIV/AIDSprevalence among South African health workers. South African Medical J2004, 94(10):846–850.67. Labonte R, Packer C, Klassen N: Managing health professional migrationfrom sub-Saharan Africa to Canada: a stakeholder inquiry into policyoptions. Human Resources for Health 2006, 4:22.68. Wadee H, Khan F: Human resources for health: health care delivery. InSouth African Health Review 2007. Edited by. Durban: Health Systems Trust;2007:141–149.69. Yassi A, Bryce E, Spiegel J: Assuming our global responsibility: Improvingworking conditions for healthcare workers globally. Open Med 2009,3(3):174–177.70. Yassi A, Bryce E, Spiegel J: Building capacity to secure healthier and saferworking conditions for healthcare workers: A South African-Canadiancollaboration. Int J Occup Environ Health 2009, 15(4):360–369.Spiegel et al. BMC Medical Informatics and Decision Making 2012, 12:84 Page 14 of 15http://www.biomedcentral.com/1472-6947/12/8471. Lavis J, Moynihan R, Oxman AD, Paulsen EJ: Evidence-informed healthpolicy 4 - Case descriptions of organizations that support the use ofresearch evidence. Implementation Sci 2008, 3:56.72. Jones D, Tremblay J, Kistnasamy B, Darwin L, Spiegel J: The implementationand evaluation of OHASIS, a health information system in a health laboratoryservice. Cancun, Mexico. Cancun, Mexico: 30thInternational Congress onOccupational Health; 2012.73. Mclaughlin J, Webster A: Rationalising knowledge: IT systems,professional identities and power. The Sociological Rev 1998,46(4):781–802.74. Sein MK, Henfridesson O, Purao S, Rossi M, Lindgren R: MIS Quarterly 2011,35:1.75. Gaydos M, Bhatia R, Morales A, Lee P, Liu S, Chang C, Salvatore A, Krause N,Minkler M: Promoting Health and Safety in San Francisco's ChinatownRestaurants: Findings and Lessons Learned from a Pilot ObservationalChecklist. Public Health Reports 2011, 126(Suppl 3):62–69.76. Comer K, Grannis S, Dixon B, Bodenhamer D, Wiehe S: IncorporatingGeospatial Capacity within Clinical Data Systems to Address SocialDeterminants of Health. Public Health Reports 2011, 126(Suppl 3):54–61.77. Beltran V, Harrison K, Hall H, Dean H: Collection of Social Determinant ofHealth Measures in U. S. National Surveillance Systems for HIV, ViralHepatitis, STDs, and TB. Public Health Reports 2011,126(Supplement 3):41–53.78. WHO, ILO, UNAIDS: Joint WHO-ILO-UNAIDS policy guidelines on improvinghealth workers’ access to HIV and tuberculosis prevention, treatment, care andsupport services. Geneva, Switzerland: WHO; 2010. Available from: http://whqlibdoc.who.int/publications/2010/9789241500692_eng.pdf79. Malik A, Khan R: Understanding the implementation of an electronic hospitalinformation system in a developing country, a case study from Pakistan.Wellington, New Zealand: 3rd Australasian Workshop on HIKM; 2009.80. Gomaa A, Occupational Healthcare Safety Network: 8th ICOH Conference.In Presentation at Occupational Health Surveillance in Health Care Seminar.Casablanca, Morocco: Occupational Healthcare Safety Network; 2010.81. Lavoie M-C, Yassi A, Bryce E, Fujii R, Longronio M, Tennassee M:International collaboration to protect health workers from infectiousdiseases. PAHO Journal 2010, 27(5):396–402.82. Pan American Health Organisation (PAHO): Regional plan on workers' healthfor the Americas. Washington: Pan American Health Organisation (PAHO):Washington; 2001.83. World Health Organisation: Workers’ health: global plan of action.Washington: World Health Organization; 2007.84. Yassi A, Bryce E, Breilh J, Lavoie M, Ndelu L, Lockhart K, Spiegel J: Researchcollaboration between infection control and occupational health inthree continents: A success story with international impact. BMC IntHealth and Human Rights 2011, 11(Suppl 2):S8.85. Spiegel J, Breilh J, Beltran E, Parra J, Solis F, Yassi A, et al: Establishing acommunity of practice of researchers, practitioners, policy-makers andcommunities to sustainably manage environmental health risks inEcuador. BMC Int Health and Human Rights 2011, 11(Suppl 2):S5.86. Observatorio Regional de Salud Colectiva Ambiente y Sociedad: PopulationHealth Observatory website. Simon Bolivar, Quito, Ecuador: University ofAndina; 2011. http://wwwuasbeduec/saludyambiente/; 2011.87. Jagger J, Perry J: Power in Numbers: Using EPINet Data to PromoteProtective Policies for Healthcare Workers. J Infusion Nursing 2002,25(6):S15–S20.88. Silva L, Hirschheim R: Fighting against windmills: Strategic informationsystems and organizational deep structures. MIS Q 2007, 31(2):327–357.89. Friesen M, Yassi A, Cooper J: Return-to-work: The importance of humaninteractions and organizational structures. Work J 2001, 17:11–22.90. Yassi A, Ostry AS, Hatter B, De Boer HM: Joint health and safety committeeeducation and the value of bipartite cooperation in the healthcaresector in British Columbia, Canada. Int J Occup Environ Health 2005,11(3):305–312.91. Ostry A, Yassi A, Ratner P, Park I, Tate R, Kidd C: Work organization andpatient care staff injuries: the impact of different care models for"alternate level of care" patients. Am J Industrial Med 2003, 44(4):392–399.92. Spiegel J, Yassi A: Occupational disease surveillance in Canada: Aframework for considering options and opportunities. Am J Industrial Med1991, 22:703–709.93. Yassi A, Tomlin K, Sidebottom C, Rideout K, De Boer H: Politics andpartnerships: challenges and rewards of partnerships in workplacehealth research in the healthcare sector of British Columbia, Canada. IntJ Occup Environ Health 2004, 10(4):457–465.94. Sandhu JS, Anderson K, Keen D, Yassi A: Implementing informationtechnology to improve workplace health: a web-based informationneeds assessment of managers in Fraser Health, British Columbia.Healthcare Manage Forum 2005, 18(4):6–16.95. Ostry A, Yassi A, Tate R: A needs assessment for occupational health andsafety programs in British Columbia's healthcare sector: A joint labour-management approach. Boston, Massachusetts: The 128th American PublicHealth Association Annual Meeting; 2000.doi:10.1186/1472-6947-12-84Cite this article as: Spiegel et al.: Tool, weapon, or white elephant? Arealist analysis of the five phases of a twenty-year programme ofoccupational health information system implementation in the healthsector. BMC Medical Informatics and Decision Making 2012 12:84.Submit your next manuscript to BioMed Centraland take full advantage of: • Convenient online submission• Thorough peer review• No space constraints or color figure charges• Immediate publication on acceptance• Inclusion in PubMed, CAS, Scopus and Google Scholar• Research which is freely available for redistributionSubmit your manuscript at www.biomedcentral.com/submitSpiegel et al. BMC Medical Informatics and Decision Making 2012, 12:84 Page 15 of 15http://www.biomedcentral.com/1472-6947/12/84


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items