research open access advancing the application of systems ... · advancing the application of...

Paina et al Health Research Policy and Systems 2014 1241httpwwwhealth-policy-systemscomcontent12141

RESEARCH Open Access

Advancing the application of systems thinkingin health exploring dual practice and itsmanagement in Kampala UgandaLigia Paina1 Sara Bennett1 Freddie Ssengooba2 and David H Peters1

Abstract

Background Many full-time Ugandan government health providers take on additional jobs ndash a phenomenon calleddual practice We describe the complex patterns that characterize the evolution of dual practice in Uganda and thelocal management practices that emerged in response in five government facilities An in-depth understanding of dualpractice can contribute to policy discussions on improving public sector performance

Methods A multiple case study design with embedded units of analysis was supplemented by interviews with policystakeholders and a review of historical and policy documents Five facility case studies captured the perspective ofdoctors nurses and health managers through semi-structured in-depth interviews A causal loop diagram illustratedinteractions and feedback between old and new actors as well as emerging roles and relationships

Results The causal loop diagram illustrated how feedback related to dual practice policy developed in Uganda Asopportunities for dual practice grew and the public health system declined over time government providersincreasingly coped through dual practice Over time government restrictions to dual practice triggered policyresistance and protest from government providers Resulting feedback contributed to compromising the supply ofgovernment providers and potentially of service delivery outcomes Informal government policies and restrictionsreplaced the formal restrictions identified in the early phases In some instances government health managersparticularly those in hospitals developed their own practices to cope with dual practice and to maintain publicsector performance Management practices varied according to the health managerrsquos attitude towards dual practice andpersonal experience with dual practice These practices were distinct in hospitals Hospitals faced challenges managinginternal dual practice opportunities such as those created by externally-funded research projects based within thehospital Private wingsrsquo inefficiencies and strict fee schedule made them undesirable work locations for providers

Conclusions Dual practice prevails because public and private sector incentives non-financial and financial arecomplementary Local management practices for dual practice have not been previously documented and providelearning opportunities to inform policy discussions Understanding how dual practice evolves and how it is managedlocally is essential for health workforce policy planning and performance discussions in Uganda and similar settings

IntroductionDual practice when health workers employed full timeby the government take on additional jobs is widespreadin developing countries particularly those with growingprivate sectors Recent studies found that 29 of physiciansin Cote drsquoIvoire 35 of physicians in Vietnam 42 inSri Lanka and 41 in Zimbabwe and as high as 80 in

Correspondence LPAINAjhuedu1Department of International Health Johns Hopkins Bloomberg School ofPublic Health 615 N Wolfe St Suite E8541 Baltimore MD 21205 USAFull list of author information is available at the end of the article

copy 2014 Paina et al licensee BioMed Central LCommons Attribution License (httpcreativecreproduction in any medium provided the orDedication waiver (httpcreativecommonsorunless otherwise stated

Indonesia and Bangladesh held second jobs [1-5] Insome contexts dual practice can be broader than privatefor-profit sector service delivery ndash including both researchand NGO work Researchers and policy-makers in devel-oping countries display increasing interest in how dualpractice affects the health system [67]Uganda is one of these countries The country has a

vibrant private health sector Within this the privatenot-for-profit health sector has for decades been actingas an extension of the public one especially after thepublic health sector was mostly destroyed during the

td This is an Open Access article distributed under the terms of the Creativeommonsorglicensesby20) which permits unrestricted use distribution andiginal work is properly credited The Creative Commons Public Domaingpublicdomainzero10) applies to the data made available in this article

Paina et al Health Research Policy and Systems 2014 1241 Page 2 of 14httpwwwhealth-policy-systemscomcontent12141

period of civil war and remains underfunded to dateUgandarsquos private for-profit sector is large fragmented anddisorganized yet very little is known about it Althoughthere are growing discussions about public-private part-nerships in health [8] dual practice seldom features onthese agendas and data on this topic is scarce In Ugandain 2005 a nationally representative survey of privatehealth facilities found that more than half (54) of privatesector doctors also declared being formally employed inthe public sector [9] While estimates from public facilitiesdo not exist in general health providers and policy-makers perceive that almost all government-employedhealth workers have dual practice In addition dual prac-tice has been rising in importance on the policy agendadue to media reports of adverse health service deliveryoutcomes [1011] as well as suspected linkages to absen-teeism and the wastage linked to it [1213] A recent studyaiming to establish policy-makersrsquo research priorities re-vealed that a principal concern was dual practice thatwas ldquoreported to greatly affect the performance of thepublic sector The dual [practice] of public healthworkers has implications on quality and management ofhealth care delivery such as indiscipline time loss andpoor work ethicsrdquo [7]Despite these concerns data on dual practice in

Uganda and elsewhere is scarce Although many types ofhealth providers are believed to engage in dual practicethe available literature examines dual practice rathernarrowly generally only from the perspective of physiciansFurthermore existing studies provide few answers toquestions related to the policy and management of dualpractice beyond agreement that the effects of dual prac-tice on the organization of the health system and servicedelivery can be either positive or negative and that theseeffects and related policy responses are highly dependenton the local context [14-16] For example if well manageddual practice may help prevent doctors from leaving thecountry by enabling them to supplement salaries withoutadversely affecting stock of doctors in the country Con-versely if poorly managed absenteeism and pilfering maynegatively affect public sector standards of care and con-tribute to inefficiencies The factors and interactions thatdrive these effects have not been explored extensivelyPresumably these factors depend on how dual practicehas evolved and how it is managed in a particular healthsystemStudying the dynamic aspects of dual practice and re-

lated interactions both within and outside the boundariesof a health system requires a departure from the lineartheoretical models found in the literature [217-20] Amore appropriate model acknowledges the holistic com-plex and adaptive nature of health systems and theirbroader environment Complex systems are composed ofmany interacting components that organize themselves in

dynamic ways are unpredictable in the long-term and areable to learn from past experiences [21-23] A researchdesign acknowledging complex systemsrsquo features as wellas potential interactions due to contextual factors is idealto guide the exploration of phenomena such as dualpractice from multiple perspectives It facilitates theexploration of complex system characteristics such asfeedback emergence and self-organization In this paperwe explore how dual practice evolved in the Ugandanhealth system and how it is currently managed in anurban environment ndash the city of Kampala with an activeprivate sector Additionally using systems approaches weattempt to reflect on why dual practice persists and thefactors underlying its current management Understandingdual practice holistically in the Ugandan context providesa basis for exploring potential policy options Gaining anin-depth understanding of the role of dual practice at vari-ous levels of the system can help policy-makers and healthmanagers to strengthen management of dual practice andits consequences

MethodsResearch designWe use case studies of urban public health facilities toinvestigate the role of dual practice and the key patternsand interactions that it motivated in the health systemReview of policy documents as well as qualitative inter-views of policy stakeholders were used to supplement aqualitative survey of workers and their managers in thestudy facilities In addition during the data analysis phasewe developed a causal loop diagram to illustrate key fac-tors and related feedback influencing dual practice in thecurrent context This paper presents only a sub-set of datathat were collected as part of a sequential exploratorymixed methods studyWe purposefully selected five public sector health facil-

ities in Kampala Uganda to represent the various levelsin the Ugandan government health system two HealthCenter III facilities one Health Center IV facility andtwo urban hospitals (see Table 1 for case characteristics)Health Center III facilities have a general outpatient clinicand a maternity ward Health Center IV is a larger facilitythan the Health Center III facilities with the capacity forinpatient services and some emergency operations Re-gional referral hospitals have specialized clinics and arestaffed by a variety of cadres including medical specialistsThe national referral hospital also has research and teach-ing components in addition to medical service provisionWithin these case studies individual respondents werepurposefully selected to ensure that at each facility theperspectives of providers (doctors and nurses) and thefacility manager were captured [24] At each facility thesample included the health facility manager (in-charge inhealth centers directors or department heads at hospitals)

Table 1 Summary of selected cases

Facility type Case A Case B Case C Case D Case E

Health Center III X X

Health Center IV X

Hospital X X

Location

Central X X X

Periphery X X

Staff composition

General practitioners X X X

Specialists X X

Nurses X X X X X

Filled positions 121 74 90 144 90

Source Ministry of Health ndash Human resources for health audit [25]Note Health Center III units are supposed to be staffed by Clinical Officersand Nurses ndash although sometimes units do have a Medical officer as well


as well as a doctor and nurse recommended by the healthfacility manager Within the larger hospital the samplingoccurred at the level of the clinical specialty [1] and there-fore included multiple manager-level respondents as wellas a nurse and a doctor recommended by each of themWithin the smaller hospital the sample included the dir-ector and two providers recommended by themPolicy stakeholders included purposefully selected indi-

viduals from professional councils relevant governmentministries private not-for-profit medical bureaus privatesector hospital administration and the local district healthoffice The main criteria guiding the sampling was theextent to which a stakeholder would know policies ondual practice either at the national level or within theirorganization would have a stake in the development of apolicy on dual practice and present a unique perspectiveon dual practice in the Ugandan context

Data collection instruments and field workA review of policy documents was undertaken before thedata collection and as these documents became availableThe main areas of interest were the existence and contentof policies actors and events that played a role in the evo-lution of dual practice in Uganda as far back in time aspossibleThe interview guides contained questions about the evo-

lution of dual practice in Uganda providersrsquo motivation toengage in dual practice advantages and disadvantages orchallenges linked to dual practice facility-level policiesand management approaches and potential policy recom-mendations Interviews with policy stakeholders focusedon policy-related questions as well as on the evolution ofdual practice in the health system Data collection tookplace during JulyndashAugust 2012 The interviews wereconducted in English Interviews were recorded unlessrespondents preferred otherwise

Data analysisAll of the recordings were transcribed and stored in Atlasti v 7 A preliminary exploratory coding structure wasconstructed based on initial readings of the transcriptsand on the conceptual framework derived from a systemsapproach to health markets and theories related to sys-tems thinking and health worker motivation [21-232627]Multiple rounds of coding focused on refining the scheme[28] During coding and analysis memos were devel-oped to capture changes in the coding structure as wellas emerging reflectionsText query results from Atlasti were arranged in matri-

ces for within and cross-case analyses according to themethods suggested by Miles and Huberman [24] For eachcase matrices were developed by theme (eg informalorganizational policies) with focus on the embedded unitsof analysis (eg summarizing and contrasting the per-spectives of health facility managers doctors and nurses)Cross-case theme analyses focused on exploring the differ-ences and similarities between the five cases specificallyby health facility type The policy stakeholder interviewswere analyzed for emerging themes along the same linesas the case studies References to the analysis and anyquotes are labeled according to respondent type to main-tain anonymity

Causal loop diagram developmentAlthough it was not an explicit goal of this study inter-views with policy-makers revealed that the role of dualpractice and the government policy on dual practice chan-ged over time and that examining this progression mightbe useful for understanding the current policy situationBased on discussions with policy-makers and case studyrespondents as well as available historical accounts wedeveloped a causal loop diagram (CLD) The CLD illus-trates the events actors and interactions ndash or the under-lying mental model and system behavior ndash that fosteredthe emergence of dual practice policy responses over timein the Ugandan health system facilitating the visualizationof complex system patterns and characteristics such aspolicy resistance feedback and adaptation [2930]The CLD was developed using Vensim PLE Plus [31]

It was challenging to recreate the history of dual prac-tice particularly in the distant past An account of themedical profession in East Africa which included de-tails about the emergence and development of dualpractice and the private sector from the perspective ofphysicians helped to identify relevant early eventsfrom the 1960s and the 1970s [32] Recent events havebeen identified from the in-depth interviews conductedfor this study and available documents The CLD wasrefined through various iterations to ensure that therelationships interactions and direction of feedbackwere most plausible


The CLD uses standard notation ldquoa positive (+) arrowfrom variable A to variable B means that A adds to B ora change in A causes a change in B in the same directiona negative (minus) arrow from A to B means that A subtractsfrom B or a change in A causes a change in B in theopposite directionrdquo [33] Some of the relationships createfeedback loops These loops are reinforcing if the vari-ables influence each other in the same direction Loopsare balancing if they influence each other in differentdirections The thickness of the line denotes researcherrsquosemphasis on a relationship for illustrative purposesDashed arrows highlight key probable relationshipsidentified through this study The question mark () indi-cates an unknown relationship This is not the recom-mended notation as it is preferred to make explicit theldquomultiple causal pathways connecting the two variablesrdquo[33] However data currently do not exist to sufficientlytease out how dual practice affects service delivery Forexample while we know that dual practice can affectsystems positively and negatively whether and how muchdual practice contributes to adverse service delivery out-comes is unknown

Ethical approvalsEthical approvals were obtained from the InstitutionalReview Board of the Johns Hopkins Bloomberg Schoolof Public Health (IRB No 4371) the Makerere UniversityCollege of Health Sciences ndash School of Public HealthHigher Degrees Research and Ethics Committee (IRBNo 11353) the Mulago Research Ethics Committee(Protocol no 249) and the Uganda National Councilfor Science and Technology (Ref No SS 2883)

ResultsTwenty-three interviews with doctors nurses and healthmanagers from various types of facilities as well as 13policy-stakeholder interviews were conducted None ofthe respondents approached for an interview declined tospeak to us although a few preferred that our interviewnot be recorded Respondent characteristics are displayedin Table 2 About half of our health facility respondentsreported having dual practice at the time of the interviewor having been previously been involved in private sectorworkThe CLDs that follow display the factors associated

with the presence of dual practice in the system and theemergence of current management practices and policiesThey illustrate three phases to describe the emergence ofdual practice in Uganda pre-independence through the1960s 1970s through the 1980s and the 1990rsquos throughthe present Table 3 complements the CLDs by illustratinga timeline of critical events that affected the policy andmanagement of dual practice

The remainder of this section first describes in eachphase the feedback and interactions that emerged in rela-tion to government policies on dual practice as well asmore broadly the development of a mixed health systemin Uganda It concludes by describing how dual practice iscurrently managed in the government facilities includedin this study

Phase 1 Dual practice policy before Ugandarsquosindependence and through the 1960sFigure 1 illustrates a relatively simple system showingno feedback or unintended consequences where a nascentprivate sector does not initially provide sufficient incen-tives for providers to engage in dual practiceDuring this time government restrictions on dual prac-

tice are formal ndash written and enforced dual practice isallowed only after government hours Some dual practiceopportunities exist however demand is low due to highsatisfaction with the government benefits Few governmentproviders chose to engage in dual practice generallyseeking the autonomy provided by private practice Noevidence was found that dual practice raised concernsabout adverse health service delivery outcomes Even as re-strictions on dual practice became stronger after Ugandarsquosindependence the Ugandan government was able to pro-vide government health workers with sufficient financialand non-financial incentives (eg satisfactory wages andthe prestige of working in a government institution re-spectively) One of the policy-stakeholder respondentsconfirms the general sentiment in this period [32]

ldquoThe assumption was that what the government payscan cater for what you require in real life [hellip] in the60rsquos a medical officer medical assistant the nursewas capable of catering for everything they requiredthe basics of life [with the government salary alone]And they were held with high esteem they were veryethical I mean a medical officer would walk with hishead highrdquo ndash Ministry of Health policy stakeholder

Phase 2 Dual practice policy in the 1970s and 1980sFigure 2 illustrates a second phase during which theUgandan system undergoes instability of military rule andeventually civil warDuring this period instability affects the health workforce

in multiple ways through reduced infrastructure andgovernment budgets as well as through persecution ofhealth providers for political reasons These country-levelhardships are intensified by the broader global recessionThe multiple crises cripple the government health systemand mark the beginning of several decades of low govern-ment salaries While the job security and prestige relatedto government service are still important the governmentfinancial incentives are no longer sufficient for providers

Table 2 Interview respondent characteristics

Facility-based respondents

Case A Case B Case C Case D1 Case E Nr ()

Gender Male 0 1 1 2 4 8 (35)

Female 3 2 2 1 7 15 (65)

Yrs in service lt10 1 1 1 0 0 3 (13)

10ndash19 0 1 1 0 5 7 (30)

20ndash29 2 0 1 1 2 6 (26)

30+ 0 1 0 2 3 6 (26)

Profession Nurse 2 2 2 1 2 9 (39)

General practitioner 0 1 1 0 0 2 (9)

Clinical officer 1 0 0 0 0 1 (4)

Specialist 0 0 0 2 9 11 (48)

Dual practice Yes 10 (43)

TOTAL 23

Policy stakeholders

Gender Male 12 (92)

Female 1 (8)

Sector Publicgovernment 5 (38)

Professional associations 4 (31)

Private for-profit 3 (23)

Private not-for-profit 1 (8)

TOTAL 131Years in service not available for one of the respondents at this facility


who remain in the system Many government providersresign at this time or leave the country Increasinglygovernment providers who remain in the system seekadditional income through dual practice The same policystakeholder explained

ldquo[With] the economic downturn of the 70s then the warsthat have been associated with [Amin and Obotersquos]regime the salary did not have any meaning anymore[hellip] The global economy has changed impactingeveryone [hellip] the country with all the hardships itrsquoshad ndash the economy has not been able to cope with themany social needs Thatrsquos why salaries across all publicservants have remained very low and therefore publicservants have to look for alternative survivalmechanismsrdquo ndash Ministry of Health policy stakeholder

As the public sector increasingly suffers and govern-ment providers ldquolook for alternative survival mechanismsrdquothis period leads to the first large-scale development ofthe private sector after Asians (including doctors) wereexpelled from Uganda for political reasons many of theUgandan government doctors who remained in the countryre-opened the former Asian private practices and many ofthem were perceived to have dual practice At this time the

increasing concerns about the quality of services providedin private for-profit medical practices (while not proven tobe linked to adverse health and health system outcomes)contributes to suspicion around dual practice specificallyrelated to potential damage to the quality of services in thepublic sector and pilfering of government medical suppliesConsequently the government begins imposing strongformal restrictions on dual practice a strict ban on dualpractice and at one point in 1972 closes all private clinicsAs shown in Figure 3 these strong formal restrictions todual practice trigger provider protests and resignationand contribute to provider migration both of which com-promise the supply of government health providers Inter-national sanctions on the military government and adeclining economy made salaries of civil savants unattract-ive Increasing protests and advocacy from professionalassociations eventually lead to the government relaxingrestrictions Weaker restrictions which allowed dual prac-tice after government hours reduced the threat to govern-ment providers and diminished the undesirable feedbackDuring the 1980s the global debt crisis and the sub-

sequent structural adjustment program fuelled thedevelopment of the private sector while at the sametime constraining government budgets [34] In this con-text the financial benefits of working in private practice

Table 3 A timeline of critical events and government policy on dual practice

Year Event Dual practice policy Consequences

Pre- Nr of African health professionalsgrowing

Weak formal govt restrictions dualpractice allowed after govt hours

None

1962 Ugandan independence

Post-1962 Govt suspicions about privatesector growing

Strong formal govt restrictions dualpractice not allowed

No immediate effects

hellipTransition to military rule andcivil war

1972 Asian doctors expulsed After 1970rsquos events restrictionsto dual practice contributed toresignations from governmentservices and provider migration ndashtherefore reducing the numberof govt providers

hellip Ugandan doctors take overprivate practices

1974 Government shuts downprivate practices

hellip Provider protest advocacy toallow dual practice

Late 1970rsquos Broadly international sanctionson military government led toeconomic collapse and declinein government salaries relativeto cost of living


Dual practice is a copingmechanism for providersremaining in Uganda

Government changes policy ondual practice as incentive forgovt providers

1980rsquos Govt suspicions about dualpractice and private sectorstrengthen

Weak formal govt restrictions dualpractice not allowed

1990rsquos

2000rsquos Rapid private sector growthespecially after system recoveredfrom civil war creates increasingnr of dual practice opportunities

No formal govt restrictions

hellip Informal govt restrictions on dualpractice with weak influence

2005ndash2007 MOH tests ban on dual practicein few hospitals

Providers threaten to resign

2009ndash2010 Office of President establishesMedicines and Health ServiceDelivery Monitoring Unit

Dual practice important copingmechanism

Increasing nr of policy discussionsaround dual practice absenteeismghost workers

Providers threaten to resign inresponse to discussions of ban

Increasing concerns about thecontribution of dual practiceto decreases in quality andaccess to care in both publicand private sectors


significantly exceeded those of the public sector and moti-vated government providers to engage in dual practiceGovernment restrictions remain formal but weak at thistime Based on the available information we propose thatduring this eventful and tumultuous period dual practiceand the incentives related to practicing in the privatesector complement incentives for government serviceMoreover restrictions on dual practice without any fur-ther measures to address the government health systemcontribute to a decrease in the number of governmentproviders The dashed lines in Figure 2 highlight theseproposed influences

Phase 3 Dual practice policy from the 1990s to thepresentFigure 4 illustrates the changes in the system from the1990s to the present the private sector grows significantlyas Uganda recovers from civil war and privatization is en-couraged through the structural adjustment program [34]and well-financed vertical health projects and clinical re-search initiativesDual practice opportunities grow quickly in a context

of rapid private sector growth as well as of increasingdonor-funded research and NGO projects generally housedwithin public facilities Due to an ever constrained budget

Figure 1 Causal loop diagram illustrating factors influencing dual practice from pre-independence through the 1960rsquos The causal loopillustrates the first period of interest a simple system with little demand for dual practice It is important to highlight that no feedback loops wereidentified in this phase


and growing population demand the government healthsystem cannot offer providers an alternative to dualpractice Private practice during this period promisessignificant financial incentives particularly in contrastto low government salaries but lacks the job security

Figure 2 Causal loop diagram illustrating factors influencing dual prasecond period of interest the health system is challenged by broader contdemand for dual practice grows so do opportunities for government provadverse effects and at first imposes a ban on dual practice A balancing lofurther details)

and prestige that are still associated with governmentpractice The increasing population demand as well as thesignificant earnings possible through private practice makedual practice a frequent coping mechanism for governmentproviders In the absence of formal rules to manage dual

ctice during the 1970rsquos and 1980rsquos The causal loop illustrates theextual events ndash such as the civil war and the global debt crisis Asiders The government becomes increasingly suspicious of potentialop first results in unintended negative consequences (See Figure 3 for

Figure 3 Focusing on the restrictions loop Starting with the 1970rsquos strong restrictions to dual practice trigger unintended consequencesthrough a balancing feedback loop ndash a decrease in the number of government providers Subsequently successful advocacy efforts to easerestrictions eventually dampen their effects on the broader health workforce although restrictions remain in place they are acceptable to theprovider population This figure re-draws the CLD diagram to better illustrate the factors influencing these unintended consequences


practice health facility managers develop their own formaland informal practices for mitigating detrimental effectsof dual practice such as absenteeism while retaining thegovernment health workforce despite low salaries andpoor infrastructureThe absence of a formal policy on dual practice was

confirmed by interview participants and also by our re-view of Ministry of Health and Ministry of Public Servicepolicy documents As our respondents illustrate belowcurrent government restrictions are informal ndash unwrittennot enforced and based on expectations of provider be-havior in the public sector

ldquoI donrsquot think there is a clear policy saying that thereis no dual practice [hellip] we are expected NOT to do it[hellip] [Health workers] know whatrsquos supposed to be thenormal but are kind of forced to do it as Irsquove said toimprove a bit on their earnings [hellip] We donrsquot comeout to fight it I canrsquot tell someone please donrsquot go theother end because therersquos a reason that is pulling thisperson to go and I have no control over that All I cando is to make sure enforcing that this person is herewith me at the right time for 7 or 8 hours So we canrsquot

influence what happens beyond that [hellip] I cannotinfluence the earnings [hellip] The person has the needsand I canrsquot satisfy the needs in any other way [hellip] Icanrsquot provide alternativesrdquo ndash Government official 1

ldquoA lot of policies are implemented while they are justknown by the policy-makers but they are not writtendown So we know about dual practice and the policyis that [hellip] it should be left as it is That people can beallowed to do dual practice [hellip] It is not written Itrsquosnot written at all but they should not take too muchof public time to do it [hellip] Unfortunately there is nomechanism to enforce how much public time peopleare going to take because [hellip] a lot of things that havegone wrong including this dual practice have gonewrong because of poor regulatory systemsrdquo ndashGovernment official 2

Interviews with policy stakeholders revealed that thegovernment initiated periodic attempts to formalizegovernment restrictions on dual practice motivated bysuspicion around dual practice due to media coverageof adverse health outcomes and poor public sector

Figure 4 Causal loop diagram illustrating factors influencing dual practice during the 1990rsquos to the present Dual practice opportunitiesgrow exponentially as it becomes more attractive to government providers working in an underfunded and over-burdened public system Aformal written government policy does not exist Local facility-level coping mechanisms emerge to mitigate negative consequences of dualpractice on the health system Periodic threats for increasing restrictions re-activate the feedback loop presented in Figure 3


performance Additionally restriction attempts are trig-gered in the context of budget discussions media reportsabout ghost workers and increased concerns about qual-ity of care in both public and private sectors as indicatedby adverse health service delivery outcomes linked toabsenteeism pilfering of drugs and patient deaths in theprivate sector from suspected malpracticeEscalating policy discussions around formalizing re-

strictions on dual practice are often met with providerprotests triggering the feedback displayed in Figure 3and the government goes back to ldquokeeping quietrdquo in thiscase meaning informal restrictions A couple of the policystakeholders provides examples of such events which alsoillustrate that the government increasingly recognizes therole of dual practice in the system particularly in theabsence of changing government pay

~2005ndash2007 ldquoThe [high level official] gave a directivethat it should stop [hellip] He said lsquoOfficer we are goingto work out the methodology of implementing it [hellip]But we shall not do it broadly across the country weshall test it in some hospitalsrsquo So we came [to one ofthe hospitals] and communicated what the [high levelofficial] had done and said these people [at thishospital] said lsquoWe hear you loud and clear but letrsquosagree if I cannot take that prescription am I free toleave the government job So that I can go to the otherside [meaning private practice] [hellip] we either stay orgo Is that what yoursquore trying to communicate to usrsquoWe said lsquoYesrsquo Within two days [hellip] the [hospital]

director came rushing to the headquarters to saylsquoGuys stop talking about dual practice becauseeverybody is winding up to gorsquo So the [governmentofficial] went back and told the [high level official]lsquoWe tried to test it in [a hospital] and all theconsultants are not bothered - they want to leave[this hospital]rsquo - and dual practice has gone onrdquo ndashMinistry of Health policy stakeholder 1

In response to the cycles of uncertainty related to in-formal government restrictions to dual practice as wellas to coping with potential negative consequences of dualpractice on public sector performance we found thatlocal facility-level and department-level managementpractices can develop in government facilities Facilitymanagement practices arise in response to increasingnumber of government providers with dual practice andaim to reduce any adverse service delivery outcomes inthe public sector As long as dual practice remains anincentive for providers to remain in government service(and sufficient resources to incentivize providers other-wise do not exist) these facility management practicescould potentially weaken any policy restrictions to dualpractice and any related negative effects on the govern-ment workforce

Local management practices for dual practiceData from the case studies revealed that in the absenceof a formal written policy on dual practice health man-agers develop their own approaches to coping with and

Paina et al Health Research Policy and Systems 2014 1241 0 of 14httpwwwhealth-policy-systemscomcontent12141

managing dual practice on a daily basis Table 4 sum-marizes the approaches identified through this studyThese facility-level management practices encourage thepresence and performance of their staff during govern-ment hours although these codes of practice are generallyunwritten For example no respondents described dualpractice being addressed directly during regular staffmeetings Instead respondents described informal one-on-one consultations health managers intervening withproviders in private often in response to an issue relatedto provider performanceIn one case a health manager fostered a culture of

flexible scheduling ie all senior doctors get one day orcertain afternoons to dedicate to their other activitieswhether research or dual practice in exchange for reportingto duty on other days According to the unitrsquos manager

ldquoWe tried to create a bit of flexibility and say ok allof us must be on station in the morning and letrsquos taketurns to cover the evening And may be trying to bringthe evening time a bit forward to to allow people toearn some extra earning [hellip] When I see the outputsthen I donrsquot complain Yes and sometimes they comeand start early before 8 orsquoclock and if someone is hereby 7 and even comes back on the weekend to clear ifthere is any backlog I think really I can only saythank you because I canrsquot pay them more than theyearnrdquo ndash Case E Health manager

This particular arrangement was not only facilitated bythe fact that the unit manager was understanding of thereasons why providers would engage in dual practiceand had an output-oriented supervision style but alsoby the fact that the majority of doctors working in thisunit worked in the same private health facility whichwas close to their government locationWithin the larger facilities formal policies included

for example having a private wing where doctors andnurses could see and get paid for private patients underthe auspices of the government facility or limiting nursesrsquonight duties so as to deter them from taking up full-timedual practice during the dayMost of the health managers interviewed had a generally

favorable attitude towards dual practice not discouragingit within their facility Their attitude stemmed from theirown personal experiences where in the past they also hadno choice but to take on additional jobs to compensate forgovernment sector shortcomings It also stemmed frombroader frustration at not being able to enforce attendancepolicies and not having the necessary tools to adequatelymonitor health workers (an exception was the healthmanager for Case C who expressed high confidence inpublic accountability) The principal tool available tomanagers for holding health workers accountable were

attendance registers which could be easily falsified Inthis context dual practice was generally toleratedwithin government facilities Health managers empha-sized the need to prioritize the completion of govern-ment duties and to the extent possible tried tointroduce incentives for improved performance in thepublic sector

ldquoI donrsquot stop anybody from doing that What [hellip] I tellthem is that priority is a core job and your core job isthe public service Once you do my work well then Idonrsquot mind about what you do nextrdquo ndash Case DNurseHealth Manager

These management approaches generally lenient withrespect to dual practice seemed to mitigate providersrsquoexit from the government health workforce Additionallythey also seemed to tackle broader issues of provider per-formance such as absenteeism In Case B providersreported being able to manage their two jobs without con-flict One of the providers reported seeing dual practice asa privilege ldquoif you want to reward yourself by doing anextra job you have to make sure we [in the governmentsector] are coveredrdquo (Case B doctor) Nevertheless becauseof the broader health system issues where managers lacktools to properly enforce policies in general approaches tomanage dual practice also had shortfalls For example inCase C the in-charge reported that providers who werefound with multiple jobs (often caught in the private fa-cilities) were asked to quit them in favor of governmentservice While this manager reported confidence in thisapproach the other respondents from the facility reportedthat almost everyone in the facility engaged in dual prac-tice but this was not discussed with the manager Some ofthese approaches also created tensions among staff Theflexible scheduling mentioned earlier was not available tonon-physicians and therefore friction arose from time totime among work teams The private wing is one of thedual practice policy interventions listed in the literature[15] however in the study context it was perceived to beinefficient and the infrastructure only marginally betterthan the rest of the facility

DiscussionThis paper is one of the few contributing empirical evi-dence on dual practice policy and management prac-tices in Uganda and low- and middle-income countriesmore broadly It illustrates how dual practice policieschanged over time in the Ugandan system and how thisphenomenon is currently managed within a sample ofgovernment facilities It also attempts to use the existingdata to reflect on and to explain why dual practice persistsand the current approaches that have developed in thestudy context

Table 4 Facility-level management practices for dual practice by caseFacility-level management practices Attitude for

dual practice1-on-1consultations

Discussion instaff meetings

Incentivessupportsupervision

Effect on the supply of government providers

Case A Dual practice allowed after governmentduties completed

Negative Yes No No Associated misunderstandings potentially createfeedback that decreases the supply of governmentproviders Providers interviewed had a differentinterpretation of the in-chargersquos version ofldquocompletenessrdquo and reported leaving governmentwork early The misunderstandings associated withthis approach were perceived to result in absenteeism

Case B Motivate providers to perform at theirpublic sector job (eg supportivesupervision tea purchased in healthmanagerrsquos personal funds)non-interference with healthworkers lives outsidegovernment duties

Cautious Yes No Yes Potentially promotes desirable feedback by creatingconditions to improve public sector performanceand retain government providers

Case C Discourage dual practice emphasizepriority for government duties and highpublic sector performance

Negative Yes No No Potentially promotes undesirable feedback byreducing the number of government providersalternatively threats of disciplinary action couldsupport improved performance in public sector

Case D Priority for government dutiesnon-interference with timeoutside government duties

Positive Yes No No Potentially does not affect government supply ofdoctors but creates tensions among staff

Although the Case D ndash the smaller hospitalrsquosleadership had a positive attitude towards dualpractice they did not report a specific managementstrategy except non-interference Doctors reportedto cope with dual practice through individualnegotiations among their colleagues however thiswas not without pitfalls as nurses were perceived tocompensate for the absence of doctors Furthermoredoctors appeared to have difficulty responding toemergencies given that they juggled two orsometimes more places of work

Case E Formal policies Mixed depends Yes No Yes in the context of flexiblescheduling NA for otherpolicies and practices

Potentially effective at reducing the number of nursesworking two full time jobs According to respondentsalso improved attendance among nurses Probably noeffect on those with part-time dual practice

Policy preventing nurses to sign up foronly night duties (which typically meansthey have a full-time day job)

A memorandum of understanding withexternally funded research projects tostop the active recruitment of governmentstaff to fill full-time positions on projects

Effective at reducing active recruitment by researchand NGO projects therefore reducing internal dualpractice opportunities According to respondentsalso improved attendance among nurses

Private wing Ineffective ndash mild effect on government providersbut has potential if more efficient

Informal policies Sustains retention among government providersparticularly specialists Flexible scheduling createsfriction among non-physiciansFlexible scheduling

Painaet

alHealth

ResearchPolicy

andSystem

s20141241

Page11

of14

httpwwwhealth-policy-system

scomcontent12141


Due to a series of health and non-health sector eventsfeedback and learning dual practice has become aninformal yet integral component of a government healthworkersrsquo incentive package This package has also evolvedover time to one where job security and prestige remainimportant but no longer sufficient due to some of thelowest salaries in the region poor system infrastructureand increasing patient loads In a situation where thegovernment cannot offer financial or non-financial al-ternatives to substitute dual practice (ie improve theincentives for sole public practice) the official policy fordual practice shifted from formal restrictions to onebased on informal expectations Any attempts to formalizerestrictions is met with unintended consequences due topolicy resistance and emerging feedback threatening thestability of government health workforce more broadlyOur data confirmed the existence of self-organization

through local facility-level management practice whichallow health managers and providers to cope with workingin both the public and private sectors The purpose ofthese practices was not necessarily to curb dual practicebut to maintain performance of the public sector byensuring the presence of providers and at the sametime to achieve an optimum balance between governmentworkersrsquo public and private activities and needs Someof these management practices were easy to identify anddescribe eg the ones guided by a health manager as inthe example of the hospital department Other manage-ment practices based on individual negotiations presum-ably depended on internal provider networks whosedevelopment and also decline could not be capturedthrough our study methods Most frequently healthmanagers found opportunities to intervene as commonsymptoms of dual practice that threaten public sectorperformance such as absenteeism triggered concernsThese management practices could potentially minimizedestabilizing effects occasionally arising from the policyfeedback and resistanceOur exploration revealed two issues that are relevant

beyond the issue of dual practice policy and managementand perhaps also beyond the Ugandan context First of allpublic sector performance management emerged as anarea with significant shortfalls In the absence of tools andsupport for rewarding good performance and punishingpoor performance the tacit indirect approach to man-aging dual practice does not sufficiently empower healthmanagers to supervise and enforce boundaries for govern-ment employees who must fulfill their duties in both thepublic and private sectors Also because the nature ofdual practice differs among nurses general practitionersand specialists cadre-specific management approachesand tools might be appropriateSecond the nature of the Ugandan health system and

that of many sub-Saharan African countries is very

different than it was immediately after independenceInitially designed around the public sector the privatesector and particularly the private for-profit componentshave been treated with suspicion and not integratedwithin a broader vision for the health system Presentlythe pluralist health systems that dominate low- andmiddle-income countries cannot be ignored As a ma-jority of the population including the poor relies on theprivate for-profit sector increasing government stew-ardship is necessary to maintain the highest standardsof service delivery [35] In this context providers engagedin dual practice could serve as a channel for reaching theprivate for-profit sector and the synergies between govern-ment practice and private for-profit practice must berecognizedHow dual practice actually affects the health system

and service delivery remains one of the key unknownsWhile the literature and study respondents acknowledgedboth positive and negative effects of dual practice inmost low- and middle-income contexts Uganda includedactual effects on the health system are unknown Further-more issues such as shortfalls in quality of care absentee-ism and efficiency gaps in public spending have broaderroot causes and can only partly be attributed to dual prac-tice A better understanding of the dual practice effects onproviders health facilities and the broad health systemwould help governments to better calibrate their policyapproach and to explore options for reaching a betterbalance between public and private sector spheres inhealth care

Strengths and limitationsThis study represents one of the few exploring dualpractice holistically from multiple perspectives (doctorsnurses managers policy stakeholders) and by applyingsystems thinking tools such as the CLD Only a fewexamples of CLDs exist in health research [36-39] Theresearchers established credibility and confirmability ofthe findings by triangulating the data from the interviewsacross multiple types of providers and where possiblethrough the document and policy reviewThe conclusions are constrained by several limitations

Much of the early history of dual practice in Uganda relieson a single source and it was not possible to verify theevents or written government documents we mention Be-cause the case studies were based in a large urban centergeneralizations to rural Uganda where the opportunitiesfor private practice are substantially different are notpossible We could not explore dual practice in privatenot-for-profit or for-profit facilities or include additionalcadres believed to engage in dual practice (eg clinicalofficers) The large hospital was much more complex thanthe other cases included in our study and perhaps deservedto be studied in greater depth Although the information


presented in the CLD was triangulated across all availabledata sources it could not be validated with study respon-dents as it was developed after the data collection endedFuture research into how dual practice is managed

by public facilities and how private for-profit facilitiesincentivize and contract with their providers would behelpful More in-depth studies looking at dual practicefrom the perspective of other cadres such as clinicalofficers or of rural practitioners could provide additionalinsights into this phenomenon The facility-level manage-ment mechanisms described here could inspire formalpolicies aimed at minimizing the negative consequencesof dual practice while helping to seize opportunities forpublic-private sector synergies The effects of dual practiceon service delivery outcomes such as quality of servicesand access to care have not been established in the litera-ture although there is consensus that dual practice likelycontributes both positively and negatively Validating theCLD and translating it into a system dynamics modelcould be relevant in policy discussions as a platform fortesting various policy scenarios and anticipating unin-tended feedback in the system

Policy relevanceThe unintended feedback revealed through the CLD attimes detrimental to the public health sector confirms therecommendations of previous studies which proposedthat a ban on dual practice would not be practical oreffective [14-1640] Periodic threats of banning dualpractice also risk destabilizing the public health sectorin places like Uganda primarily by reducing the supplyof government health workers The private for-profitsector allows government providers the additional finan-cial resources that the Ugandan government is currentlynot able to supply In a relationship of mutual depend-ency government providers in dual practice allow for thegrowth of the private for-profit sector in the context oflimited health workforce and increasing populationdemandIn the short term the Ugandan government should

consider the promotion of policies that are flexible tolocal adaptions to promote access and quality of servicesin the public sector while at the same time allowing suf-ficient income for government providers Informal localadaptations to managing dual practice exist in Ugandaand provide a natural experiment for various dual prac-tice policies In the long term the Ugandan governmentshould consider broader improvements to public sectormanagement and increasing the resources available tothe health sector as well as increasing synergies with theprivate sectorReforms currently under discussion in Uganda include

health insurance and performance-based contracts bothwould change how providers are paid Such reforms could

potentially provide an entry point for strengthening publicsector management in general and therefore providehealth facility managers the tools they are currentlylacking to manage dual practice As dual practice is un-likely to disappear in the short term its existence androle in the health sector cannot be ignored during thedesign and implementation of major health reforms inUganda and other countries where dual practice existsPotential unintended effects (feedback) should be an-ticipated based on past events related to dual practiceand dealt with accordingly

Competing interestsThe authors declare that they have no competing interests

Authorsrsquo contributionsThis study was part of LPrsquos doctoral dissertation DP SB and FS were part ofthe advising and thesis committees making important contributions to allphases of design implementation and analysis LP prepared the first draft ofthe manuscript DP SB and FS contributed to revisions and finalizing themanuscript All authors read and approved the final manuscript

AcknowledgementsThis paper is part of the Thematic Series entitled ldquoAdvancing the applicationof systems thinking in healthrdquo The Series was coordinated by the Alliance forHealth Policy and Systems Research World Health Organization Thepublication of the Series and the associated capacity building anddissemination activities were carried out with the aid of a grant from theInternational Development Research Centre Ottawa CanadaThe authors would also like to give thanks to Ms Taghreed Adam and to thetwo reviewers of the paper for their insightful suggestionsSupport for Sara Bennett and David Peters was provided through the FutureHealth Systems Research Programme Consortium funded by the UnitedKingdomrsquos Department for International Development (DFID)

Author details1Department of International Health Johns Hopkins Bloomberg School ofPublic Health 615 N Wolfe St Suite E8541 Baltimore MD 21205 USA2Department of Health Policy Planning and Management School of PublicHealth College of Health Sciences Makerere University Mulago Hill Rd POBox 7072 Kampala Uganda

Received 6 January 2014 Accepted 10 June 2014Published 18 August 2014

References1 Gruen R Anwar R Begum T Killingsworth JR Normand C Dual job holding

practitioners in Bangladesh an exploration Soc Sci Med 2002 54267ndash2792 Berman P Cuizon D Multiple public-private jobholding of health care

providers in developing countries an exploration of theory and evidenceIn Multiple Public-Private Jobholding of Health Care Providers in DevelopingCountries An Exploration of Theory and Evidence London Department forInternational Development ndash Health Systems Resource Centre 2004

3 Chomitz K Setiadi G Azwar A Ismail N Widiyarti O What do doctorswant Developing incentives for doctors to serve in Indonesiarsquos rural andremote areas In What do Doctors Want Developing Incentives for Doctors toServe in Indonesiarsquos Rural and Remote Areas Washington DC WorldBank ndash Development Research Group 1998

4 Gupta N Dal Poz MR Assessment of human resources for health usingcross-national comparison of facility surveys in six countries Hum ResourHealth 2009 722

5 Vujicic M Shengelia B Alfano M Thu HB Physician shortages in ruralVietnam Using a labor market approach to inform policy Soc Sci Med2011 73(7)970ndash977

6 Ranson MK Chopra M Atkins S Dal Poz MR Bennett S Priorities forresearch into human resources for health in low- and middle-incomecountries Bull World Health Organ 2010 88435ndash443


7 Asiimwe D Identification of priority research questions within the areasof health financing human resources for health and the role of non-statesector In Identification of Priority Research Questions within the Areas ofHealth Financing Human Resources for Health and the Role of Non-StateSector Makerere Makerere Institute of Social Research 2008

8 Government of Uganda ndash Ministry of Health National policy on publicprivate partnership in health In National Policy on Public Private Partnershipin Health Kampala Ministry of Health 2011

9 Mandelli A Kyomuhangi LB Scribner S Survey of Private Health Facilitiesin Uganda In Survey of Private Health Facilities in Uganda Bethesda MDPartners for Health Reform Plus (PHRplus) ndash Abt Associates Inc 2005

10 Karugaba M Kwesiga P Doctor suspended over failure to attend to patientKampala Uganda The Vision Group New Vision - Ugandas LeadingDaily 2011

11 Kiwawulo C Nsubuga H Nakamya went to hospital to give life life becamedeath Kampala Uganda The Vision Group New Vision - Ugandas LeadingDaily 2013

12 Chaudhury N Hammer J Kremer M Muralidharan K Rogers FH Missing inaction teacher and health worker absence in developing countriesJ Econ Perspect 2006 2091ndash116

13 Okwero P Tandon A Sparkes S McLaughlin J Hoogeveen J Fiscal Space forHealth in Uganda Washington DC The World Bank 2010

14 Socha KZ Bech M Physician dual practice a review of literature HealthPolicy 2011 1021ndash7

15 Kiwanuka SN Rutebemberwa E Nalwadda C Okui O Ssengooba FKinengyere AA Pariyo GW Interventions to manage dual practice amonghealth workers Cochrane Database Syst Rev 2011 CD008405

16 Ferrinho P Lerberghe W Fronteira I Hipoacutelito F Biscaia A Dual practice inthe health sector review of the evidence Hum Resour Health 200421ndash17

17 Eggleston K Bir A Physician dual practice Health Policy 2006 78157ndash16618 Biglaiser G Ma C-tA Moonlighting public service and private practice

RAND J Econ 2007 382119 Gonzalez P Should physiciansrsquo dual practice be limited An incentive

approach Health Econ 2004 13505ndash52420 Gonzalez P Macho-Stadler I A theoretical approach to dual practice

regulations in the health sector J Health Econ 2013 3266ndash8721 Paina L Peters DH Understanding pathways for scaling up health

services through the lens of complex adaptive systems Health Policy Plan2012 27(5)365ndash373

22 Adam T de Savigny D Systems thinking for strengthening healthsystems in LMICs need for a paradigm shift Health Policy Plan 201227iv1ndashiv3

23 Tan J Wen JH Awad N Health Care and Service Delivery Systems asComplex Adaptive Systems Commun ACM 2005 48(5)36ndash44

24 Miles MB Huberman AM Saldantildea J Qualitative Data Analysis A MethodsSourcebook Thousand Oaks CA SAGE Publications Inc 2013

25 Ministry of Health Human Resources for Health Audit Report KampalaUganda Ministry of Health 2009

26 Bloom G Champion C Lucas H Peters D Standing H Making healthmarkets work better for poor people Improving provider performanceBaltimore MD Future Health Systems 2009 httpwwwihf-fihorgcontentdownload4493433fileMaking20health20markets20work20better20for20poor20people20improving20provider20performancepdf

27 Franco LM Bennett S Kanfer R Health sector reform and public sectorhealth worker motivation a conceptual framework Soc Sci Med 2002541255ndash1266

28 Saldantildea J The Coding Manual for Qualitative Researchers Thousand Oaks CASAGE Publications Inc 2009

29 Forrester JW Industrial Dynamics Cambridge MA MIT Press 196130 Sterman JD Learning from evidence in a complex world Am J Public

Health 2006 96505ndash51431 Vensim Personal Learning Edition (2012) from httpvensimcom Accessed

December 1 2013 httpvensimcomvensim-personal-learning-edition32 Iliffe J East African Doctors A History of the Modern Profession 2nd edition

Kampala Uganda Fountain Publishers 199833 Sterman JD Business Dynamics Systems Thinking and Modeling for a

Complex World New York NY McGraw Hill 200034 Opio F The Impact of Structural Adjustment of Poverty and Income

Distribution in Uganda Makerere Economic Policy Research CentreMakerere University 1996

35 International Finance Corporation The Business of Health in Africa Partneringwith the Private Sector to Improve Peoplersquos Lives Washington DC IFC WorldBank Group 2008

36 Rwashana AS Williams DW Neema S System dynamics approach toimmunization healthcare issues in developing countries a case study ofUganda Health Informatics J 2009 1595ndash107

37 Agyepong IA Kodua A Adjei S Adam T When lsquosolutions of yesterdaybecome problems of todayrsquo crisis-ridden decision making in a complexadaptive system (CAS)mdashthe Additional Duty Hours Allowance in GhanaHealth Policy Plan 2012 27iv20ndashiv31

38 Rwashana AS Williams DW Modeling the dynamics of immunizationhealthcare systems - the Ugandan case study In The 26th InternationalConference of the System Dynamics Society July 20-July 24 Athens Greece2008

39 Patel B Chaussalet T Millard P Balancing the NHS balanced scorecard EurJ Oper Res 2008 185905ndash914

40 Garciacutea-Prado A Gonzaacutelez P Policy and regulatory responses to dualpractice in the health sector Health Policy 2007 84142ndash152

doi1011861478-4505-12-41Cite this article as Paina et al Advancing the application of systemsthinking in health exploring dual practice and its management inKampala Uganda Health Research Policy and Systems 2014 1241

Submit your next manuscript to BioMed Centraland take full advantage of

bull Convenient online submission

bull Thorough peer review

bull No space constraints or color figure charges

bull Immediate publication on acceptance

bull Inclusion in PubMed CAS Scopus and Google Scholar

bull Research which is freely available for redistribution

Submit your manuscript at wwwbiomedcentralcomsubmit

Abstract

Background

Methods

Results

Conclusions

Introduction

Methods

Research design

Data collection instruments and field work

Data analysis

Causal loop diagram development

Ethical approvals

Results

Phase 1 Dual practice policy before Ugandarsquos independence and through the 1960s

Phase 2 Dual practice policy in the 1970s and 1980s

Phase 3 Dual practice policy from the 1990s to the present

Local management practices for dual practice

Discussion

Strengths and limitations

Policy relevance

Competing interests

Authorsrsquo contributions

Acknowledgements

Author details

References

Paina et al Health Research Policy and Systems 2014 1241 of 14httpwwwhealth-policy-systemscomcontent12141

period of civil war and remains underfunded to dateUgandarsquos private for-profit sector is large fragmented anddisorganized yet very little is known about it Althoughthere are growing discussions about public-private part-nerships in health [8] dual practice seldom features onthese agendas and data on this topic is scarce In Ugandain 2005 a nationally representative survey of privatehealth facilities found that more than half (54) of privatesector doctors also declared being formally employed inthe public sector [9] While estimates from public facilitiesdo not exist in general health providers and policy-makers perceive that almost all government-employedhealth workers have dual practice In addition dual prac-tice has been rising in importance on the policy agendadue to media reports of adverse health service deliveryoutcomes [1011] as well as suspected linkages to absen-teeism and the wastage linked to it [1213] A recent studyaiming to establish policy-makersrsquo research priorities re-vealed that a principal concern was dual practice thatwas ldquoreported to greatly affect the performance of thepublic sector The dual [practice] of public healthworkers has implications on quality and management ofhealth care delivery such as indiscipline time loss andpoor work ethicsrdquo [7]Despite these concerns data on dual practice in

Uganda and elsewhere is scarce Although many types ofhealth providers are believed to engage in dual practicethe available literature examines dual practice rathernarrowly generally only from the perspective of physiciansFurthermore existing studies provide few answers toquestions related to the policy and management of dualpractice beyond agreement that the effects of dual prac-tice on the organization of the health system and servicedelivery can be either positive or negative and that theseeffects and related policy responses are highly dependenton the local context [14-16] For example if well manageddual practice may help prevent doctors from leaving thecountry by enabling them to supplement salaries withoutadversely affecting stock of doctors in the country Con-versely if poorly managed absenteeism and pilfering maynegatively affect public sector standards of care and con-tribute to inefficiencies The factors and interactions thatdrive these effects have not been explored extensivelyPresumably these factors depend on how dual practicehas evolved and how it is managed in a particular healthsystemStudying the dynamic aspects of dual practice and re-

lated interactions both within and outside the boundariesof a health system requires a departure from the lineartheoretical models found in the literature [217-20] Amore appropriate model acknowledges the holistic com-plex and adaptive nature of health systems and theirbroader environment Complex systems are composed ofmany interacting components that organize themselves in

dynamic ways are unpredictable in the long-term and areable to learn from past experiences [21-23] A researchdesign acknowledging complex systemsrsquo features as wellas potential interactions due to contextual factors is idealto guide the exploration of phenomena such as dualpractice from multiple perspectives It facilitates theexploration of complex system characteristics such asfeedback emergence and self-organization In this paperwe explore how dual practice evolved in the Ugandanhealth system and how it is currently managed in anurban environment ndash the city of Kampala with an activeprivate sector Additionally using systems approaches weattempt to reflect on why dual practice persists and thefactors underlying its current management Understandingdual practice holistically in the Ugandan context providesa basis for exploring potential policy options Gaining anin-depth understanding of the role of dual practice at vari-ous levels of the system can help policy-makers and healthmanagers to strengthen management of dual practice andits consequences

MethodsResearch designWe use case studies of urban public health facilities toinvestigate the role of dual practice and the key patternsand interactions that it motivated in the health systemReview of policy documents as well as qualitative inter-views of policy stakeholders were used to supplement aqualitative survey of workers and their managers in thestudy facilities In addition during the data analysis phasewe developed a causal loop diagram to illustrate key fac-tors and related feedback influencing dual practice in thecurrent context This paper presents only a sub-set of datathat were collected as part of a sequential exploratorymixed methods studyWe purposefully selected five public sector health facil-

ities in Kampala Uganda to represent the various levelsin the Ugandan government health system two HealthCenter III facilities one Health Center IV facility andtwo urban hospitals (see Table 1 for case characteristics)Health Center III facilities have a general outpatient clinicand a maternity ward Health Center IV is a larger facilitythan the Health Center III facilities with the capacity forinpatient services and some emergency operations Re-gional referral hospitals have specialized clinics and arestaffed by a variety of cadres including medical specialistsThe national referral hospital also has research and teach-ing components in addition to medical service provisionWithin these case studies individual respondents werepurposefully selected to ensure that at each facility theperspectives of providers (doctors and nurses) and thefacility manager were captured [24] At each facility thesample included the health facility manager (in-charge inhealth centers directors or department heads at hospitals)




Health Center IV X

Hospital X X

Location

Central X X X

Periphery X X

Staff composition


Specialists X X

Nurses X X X X X


























Gender Male 0 1 1 2 4 8 (35)

Female 3 2 2 1 7 15 (65)

Yrs in service lt10 1 1 1 0 0 3 (13)

10ndash19 0 1 1 0 5 7 (30)

20ndash29 2 0 1 1 2 6 (26)

30+ 0 1 0 2 3 6 (26)




Specialist 0 0 0 2 9 11 (48)


TOTAL 23

Policy stakeholders

Gender Male 12 (92)

Female 1 (8)
















None
















1990rsquos





































































Painaet

alHealth

ResearchPolicy

andSystem

s20141241

Page11

of14


scomcontent12141









































































Abstract

Background

Methods

Results

Conclusions

Introduction

Methods

Research design


Data analysis


Ethical approvals

Results





Discussion


Policy relevance

Competing interests


Acknowledgements

Author details

References




Health Center IV X

Hospital X X

Location

Central X X X

Periphery X X

Staff composition


Specialists X X

Nurses X X X X X


























Gender Male 0 1 1 2 4 8 (35)

Female 3 2 2 1 7 15 (65)

Yrs in service lt10 1 1 1 0 0 3 (13)

10ndash19 0 1 1 0 5 7 (30)

20ndash29 2 0 1 1 2 6 (26)

30+ 0 1 0 2 3 6 (26)




Specialist 0 0 0 2 9 11 (48)


TOTAL 23

Policy stakeholders

Gender Male 12 (92)

Female 1 (8)
















None
















1990rsquos





































































Painaet

alHealth

ResearchPolicy

andSystem

s20141241

Page11

of14


scomcontent12141









































































Abstract

Background

Methods

Results

Conclusions

Introduction

Methods

Research design


Data analysis


Ethical approvals

Results





Discussion


Policy relevance

Competing interests


Acknowledgements

Author details

References















Gender Male 0 1 1 2 4 8 (35)

Female 3 2 2 1 7 15 (65)

Yrs in service lt10 1 1 1 0 0 3 (13)

10ndash19 0 1 1 0 5 7 (30)

20ndash29 2 0 1 1 2 6 (26)

30+ 0 1 0 2 3 6 (26)




Specialist 0 0 0 2 9 11 (48)


TOTAL 23

Policy stakeholders

Gender Male 12 (92)

Female 1 (8)
















None
















1990rsquos





































































Painaet

alHealth

ResearchPolicy

andSystem

s20141241

Page11

of14


scomcontent12141









































































Abstract

Background

Methods

Results

Conclusions

Introduction

Methods

Research design


Data analysis


Ethical approvals

Results





Discussion


Policy relevance

Competing interests


Acknowledgements

Author details

References





None
















1990rsquos





































































Painaet

alHealth

ResearchPolicy

andSystem

s20141241

Page11

of14


scomcontent12141









































































Abstract

Background

Methods

Results

Conclusions

Introduction

Methods

Research design


Data analysis


Ethical approvals

Results





Discussion


Policy relevance

Competing interests


Acknowledgements

Author details

References























































Painaet

alHealth

ResearchPolicy

andSystem

s20141241

Page11

of14


scomcontent12141









































































Abstract

Background

Methods

Results

Conclusions

Introduction

Methods

Research design


Data analysis


Ethical approvals

Results





Discussion


Policy relevance

Competing interests


Acknowledgements

Author details

References









































































Abstract

Background

Methods

Results

Conclusions

Introduction

Methods

Research design


Data analysis


Ethical approvals

Results





Discussion


Policy relevance

Competing interests


Acknowledgements

Author details

References

research open access advancing the application of systems ... · advancing the application of...

Documents