THE PROBLEM WITH The problem with ... - BMJ Quality & …

1 The PROBLEM with Plan-Do-Study-ActcyclesJulie E Reed,1 Alan J Card2,31 NIHR CLAHRC NWL, ImperialCollege London, London, UK2 Department of Management,University of Notre Dame, NotreDame, Indiana, USA3 Evidence-Based HealthSolutions, LLC, Notre Dame,Indiana, USAC orrespondence toDr Julie E Reed, NIHR CLAHRCNWL, Imperial College London,Chelsea and WestminsterHospital, 369 Fulham Road,London SW10 9NH, 23 November 2015 Published Online First23 December 2015 cite:Reed JE, Card Qual Saf2016;25:147 improvement (QI) methods havebeen introduced to healthcare to supportthe delivery of care that is safe, timely,effective, efficient, equitable and costeffective. Of the many QI tools andmethods, the Plan-Do-Study-Act (PDSA)cycle is one of the few that focuses onthe crux of change, the translation ofideas and intentions into action. As such,the PDSA cycle and the concept of itera-tive tests of change are central to manyQI approaches, including the model forimprovement,1lean,2six sigma3and totalquality provides a structured experimen-tal learning approach to testing , concerns have been raisedregarding the fidelity of application ofPDSA method, which may underminelearning efforts,5the complexity of itsuse in practice56and as to the appropri-ateness of the PDSA method to addressthe significant challenges of article presents our reflections onthe full potential of using PDSA inhealthcare, but in doing so we explorethe inherent complexity and multiplechallenges of executing PDSA.

2 We argue that the problemwith PDSA is the oversimplification ofthe method as it has been translated intohealthcare and the failure to invest in arigorous and tailored application of VALUE OF PDSA IN HEALTHCAREIMPROVEMENTThe purpose of the PDSA method lies inlearning as quickly as possible whether anintervention works in a particular settingand to making adjustments accordingly toincrease the chances of delivering andsustaining the desired improvement. Incontrast to controlled trials, PDSA sallow new learning to be built in to thisexperimental process. If problems areidentified with the original plan, then thetheory can be revised to build on thislearning and a subsequent experimentconducted to see if it has resolved theproblem, and to identify if any furtherproblems also need to be addressed. Inthe complex social systems of healthcare,this flexibility and adaptability of PDSAare important features that support theadaption of interventions to work inlocal successful PDSA process does notequal a successful QI project or pro-gramme.

3 The intended output of PDSA islearning and informed action. Successfulapplication of the PDSA methodologymay enable users to achieve their QIgoals more efficiently or to reach QIgoals they would otherwise not haveachieved. But it isalsosuccessful if itsaves wasted effort by revealing QI goalsthatcannotbe achieved under realisticconstraints or if it identifies new pro-blems to tackle instead of the originallyidentified issue. A well-conducted PDSA promises learning. But it does not, andcannot, promise that users will achievetheir desired PDSA has been translated intohealthcare from industrial settings, anemphasis has been placed on rapidsmall-scale tests of change, often on one,three and then five patients in ramps ofincreasing scale, and responsibility dele-gated to frontline staff and improvementor Quality managers. This pragmaticapproach has been embraced and hasbeen seen as providing a new freedomfor healthcare staff to lead change andimprovement in local care , the process of change rarelyprogresses in simple linear of PDSAs can reveal otherrelated issues that need to be addressed inorder to achieve the improvement issues may relate to minor changesto current practices or processes of care,but can often reveal larger cultural ororganisational issues that need to beaddressed and PROBLEM JE, Card Qual Saf2016;25:147 152.

4 On July 7, 2022 by guest. Protected by Qual Saf: first published as on 23 December 2015. Downloaded from on July 7, 2022 by guest. Protected by Qual Saf: first published as on 23 December 2015. Downloaded from on July 7, 2022 by guest. Protected by Qual Saf: first published as on 23 December 2015. Downloaded from Recent evaluations have reported on the failure ofthe PDSA method to help frontline staff address themultiple improvement challenges they faced as thescale of investigation and range of issues they neededto address report evaluating the SaferClinical Systems programme in the UK identified theneed for clarity about when improvement approachesbased on PDSA cycles are appropriate and when theyare not , viewing some challenges as too big andhairy for the PDSA method and beyond the scope ofsmall-scale tests of change run by local clinical argue that any improvement situation, no matterhow big and hairy, is conducive to application of thePDSA method. The four stages of PDSA mirror thescientific experimental method of formulating ahypothesis, collecting data to test this hypothesis, ana-lysing and interpreting the results and making infer-ences to iterate the improvement initiatives have been plannedat national level to support standardisation of care orplanned over a cup of coffee to solve a minor localproblem, we believe there will always be a role forPDSA.

5 In moving from planning to implementing achange in practice, PDSA provides a structure forexperimental learning to know whether a change hasworked or not, and to learn and act upon any newinformation as a it is not a magic bullet. Increasingly complexproblems require increasingly sophisticated applicationof the PDSA method, and this is where we believe theproblem with the PDSA method SIMPLICITY BELIES ITS SOPHISTICATIONOne of the main narratives surrounding the use ofPDSA in healthcare is that it is easy, and can beapplied in practice by anyone. At one level this is true,and the simplicity of the PDSA method and its applic-ability to many different situations can be viewed asone of its main strengths. However, this simplicityalso creates some of the greatest challenges to usingPDSA successfully. Users need to understand how toadapt the use of PDSA to address different problemsand different stages in the lifecycle of each improve-ment project. This requires an extensive repertoire ofskills and knowledge to be used in conjunction withthe basic PDSA of the main problems encountered in usingPDSA is the misperception that it can be used as astandalone method.

6 PDSA needs to be used as part ofa suite of QI methods, the exact nature of which maybe influenced by the broader methodologicalapproach that is being followed (eg, model forimprovement, lean ). An important role of the widermethodological approach is to conduct investigationsprior to starting the use of PDSA to ensure that theproblem is correctly understood and can include process mapping, failuremode effects analysis, cause and effect analysis,stakeholder engagement and interviews, data analysisand review of existing second misperception is that the PDSA is limitedto small-scale tests of change on one, three and fivepatients. PDSA is an extremely flexible method thatcan be adapted to support the scale up of interven-tions and used in conjunction with monitoring activ-ities to support sustainability. But, this flexibility givesrise to a number of key dimensions that requirecareful consideration. For instance, the scope andscale of change, the amount of preparation prior touse, rigour of the evaluation, time, expertise, manage-ment support and funding must be carefully these needs must be rebalanced over the pro-ject s lifecycle.

7 If managed well, these adjustmentsenable the use of PDSA to adapt to new learning andsupport the design and conduct of tests of change asthey increase in scale, and often complexity, toachieve the desired improvement PDSA as an iterative design framework tohelp solve big hairy problems or big hairy audaciousgoals 11is, therefore, entirely appropriate. In fact,developing solutions to large-scale wicked pro-blems 12mayrequire an iterative explorative and gen-erative 13approach of the sort PDSA provides, inwhich knowledge is built through designing .13 Thekey is to understand that this framework will need tobe implemented (and resourced) very differently forlarge and complex problems than for smaller andmore tame problems. One size does not fit frontline staff with little training or supportmay successfully address some Quality problems, thecomplexity of many problems demands greater organ-isational support, with direct involvement of seniormanagers to facilitate adequate planning.

8 Projects inwhich frontline staff must fend for themselves alsorun the risk of insufficient usage of theory and exist-ing evidence to develop the intervention and a sub-optimal (NOT DIRTY) TESTS OF CHANGEIn healthcare, PDSA training often overemphasises theconceptual simplicity of the framework and underem-phasises the different ways in which the method canbe adapted to solve increasingly complex frequently leads people to leap into PDSA withinsufficient prior investigation and framing of theproblem, to delegate management of the process tofrontline staff who have little influence over broadersystemic concerns that need to be addressed, and toprovide these staff with little support to overcome theobstacles and barriers they face. The resources, skillsand expertise required to apply PDSA in the realworld are often significantly underestimated, leadingto projects that are destined to has led to the impression that PDSA cyclesinvolve quick and dirty tests of change. In the rushto empower healthcare staff, there is a danger that theThe PROBLEM JE, Card Qual Saf2016;25:147 152.

9 On July 7, 2022 by guest. Protected by Qual Saf: first published as on 23 December 2015. Downloaded from scientific rigour of the PDSA method is frequentlycompromised. A systematic review5revealed that thecore principles of PDSA are often not executed inpractice, with substantial variability with which theyare designed, executed and reported in the healthcareliterature .6A failure to properly execute PDSAs canundermine learning if data collection doesnot occur frequently enough, if iterative cycles arefew, and if system-level changes are not apparent as aresult of these cycles, the improvement work is lesslikely to succeed .6 While its scientific principles differfrom those of controlled trials, rigour in the applica-tion of PDSA is still required for PDSA to maximisethe learning obtained from tests of addition to a lack of fidelity with PDSA guidingprinciples, there is the need to ensure that each stageof the cycle is conducted well. But the frenetic cultureendemic in healthcare organisations can make it diffi-cult to achieve sustained engagement in the delibera-tive processes of GET ON with ITWhile planning paralysis can be an issue in health-care organisations, the more common PROBLEM is aserious underinvestment in the planning phase.

10 Thepervasive cultural compulsion to just get on with it 14leads many teams to move too quickly from plan to do. The consequences of skipping this up-front workcan include wasted PDSA cycles or projects that 1describes some of the key failuremodes for the planning and preplanning (ie, investiga-tion and PROBLEM -framing) steps of the PDSA do planning failures present such a challengeto the successful use of PDSA? It is much more diffi-cult to correctly execute and learn from a plan thathas not been well thought out. And even perfect exe-cution cannot ensure success if the plan, itself, iterative nature of PDSA enables course correc-tions, but this feature of the approach is much moreeffective if there was a clear and reasoned course inTable 1 Key failure modes for the investigation/ PROBLEM framing and plan stepsPDSA stagesKey failure modesPotential consequenceInvestigation and PROBLEM framingDefine the PROBLEM ; determine its causes/contributing factors; identify stakeholders; setthe criteria for successPoor definition of the PROBLEM and its causes/contributing factors1521 25 Time, money and goodwill may be wasted trying tosolve the wrong PROBLEM or solve it in the wrongwayFailure to clearly define the criteria for success andhow performance will be measured52226A poor match between the design of theintervention and its intended impact.