Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Applying the quality improvement collaborative method to process redesign: a multiple case study

Applying the quality improvement collaborative method to process redesign: a multiple case study Background: Despite the widespread use of quality improvement collaboratives (QICs), evidence underlying this method is limited. A QIC is a method for testing and implementing evidence-based changes quickly across organisations. To extend the knowledge about conditions under which QICs can be used, we explored in this study the applicability of the QIC method for process redesign. Methods: We evaluated a Dutch process redesign collaborative of seventeen project teams using a multiple case study design. The goals of this collaborative were to reduce the time between the first visit to the outpatient’s clinic and the start of treatment and to reduce the in-hospital length of stay by 30% for involved patient groups. Data were gathered using qualitative methods, such as document analysis, questionnaires, semi-structured interviews and participation in collaborative meetings. Results: Application of the QIC method to process redesign proved to be difficult. First, project teams did not use the provided standard change ideas, because of their need for customised solutions that fitted with context- specific causes of waiting times and delays. Second, project teams were not capable of testing change ideas within short time frames due to: the need for tailoring changes ideas and the complexity of aligning interests of involved departments; small volumes of involved patient groups; and inadequate information and communication technology (ICT) support. Third, project teams did not experience peer stimulus because they saw few similarities between their projects, rarely shared experiences, and did not demonstrate competitive behaviour. Besides, a number of project teams reported that organisational and external change agent support was limited. Conclusions: This study showed that the perceived need for tailoring standard change ideas to local contexts and the complexity of aligning interests of involved departments hampered the use of the QIC method for process redesign. We cannot determine whether the QIC method would have been appropriate for process redesign. Peer stimulus was non-optimal as a result of the selection process for participation of project teams by the external change agent. In conclusion, project teams felt that necessary preconditions for successful use of the QIC method were lacking. Background based practice through sharing knowledge with others Quality improvement collaboratives (QICs) are used in a similar setting over a short period of time [4]. increasingly in many countries to achieve large-scale Within the QIC method, external change agents provide improvements in performance and to provide specific collaborative project teams from different healthcare remedies to overcome the typically slow diffusion of departments or organisations with a clear vision for medical and healthcare innovations [1-3]. A QIC is a ideal care in the topic area and a set of specific changes multifaceted method that seeks to implement evidence- that may improve system performance significantly [5,6]. Project teams also learn from the external change agent about the model for improvement. The model for * Correspondence: [email protected] improvement incorporates four key elements [6]: speci- NIVEL, Netherlands Institute for Health Services Research, P.O. Box 1568, fic and measurable aims; measures of improvement that 3500 BN Utrecht, the Netherlands © 2010 Vos et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. A P S D A P S D A P S D A P S D A P A P S D A P S D Vos et al. Implementation Science 2010, 5:19 Page 2 of 11 http://www.implementationscience.com/content/5/1/19 are tracked over time; key changes that will result in the In addition to the relatively efficient use of external desired improvement; and series of parallel testing plan- change agent support and the exchange of change ideas do-study-act (PDSA) cycles. Each series involves a test as well as the model for improvement, the strength of of onechangeidea(Figure1, part A) [7]. On thebasis the QIC method seems to be that collaborative project of the results of the first test of one series, a project teams share experiences of making changes, which team can decide to refine the change idea (in case the accelerates the rate of improvement (peer stimulus) [3]. change idea works in their context) or to start a new However, despite the widespread use of QICs, a recent test series of a new change idea (in case the test did not review on their impact indicates that evidence is positive lead to the desired result). These PDSA cycles should be but limited, and the effects cannot be predicted with cer- short but significant, testing a big change idea in a short tainty [5]. This apparent inconsistency requires a deeper timeframe so that a team can identify ways to improve understanding of how and why QICs work. Therefore it or change theidea[8].InFigure2,anexampleisgiven is necessary to explore the ‘black box’ of the intervention to illustrate the model for improvement. and to study the determinants of success or failure of the A. Testing and implementing changes according to the QIC methodology Example of a series of linked testing cycles Modify the protocol and Reengineered system make it standard practice Use the entire protocol with all patients T4 Modify the protocol and use it with other patients T3 Use part of a protocol with small group of T2 patients and refine it T1 Idea 1: Concept Design: Idea 2 Idea 3 Idea 4 one-stop-shop Source original figure: Langley GJ et al . (1996) [5] B. Testing and implementing changes according to the advised method in the evaluated collaborative Part a of Part b of Part c of Part d of Idea 1: One-stop-shop protocol protocol protocol protocol (e.g. triage) (e.g. planning all diagnostics in one day) Part a , Part b , Part c, Part d , Idea 2 Idea 2 Idea 2 Idea 2 Idea 2 Tekst Part a , Part b , Part c, Part d , Idea 3 Idea 3 Idea 3 Idea 3 Idea 3 Part a , Part b , Part c, Part d , Idea 4 Idea 4 Idea 4 Idea 4 Idea 4 Figure 1 Testing and implementing changes using PDSA cycles A P S D esig Detail D n A P A S D S D A P S D A P S D A P S D Concept design Reengineered system A P S D A P S D Vos et al. Implementation Science 2010, 5:19 Page 3 of 11 http://www.implementationscience.com/content/5/1/19 1. Aim --------------------------------------------------- Figure 2 Applying the model for improvement, an example QIC method [5,9]. In this article, we contribute to this by Beter’ (‘Better Faster’), which began in 2004 as an initia- assessing the applicability of this quality improvement tive from the Ministry of Health and the Dutch Hospital method to process redesign. Process redesign aims to Association. ‘Sneller Better’ aimed to realise substantial improve the organisation of care delivery in terms of and appealing performance improvements in three waiting times in a patients’ care trajectory. From other groups of eight Dutch hospitals in the areas of patient studies it is already known that the QIC method can be logistics and safety. These twenty-four hospitals were successfully applied to improve the organisation of care enrolled in the programme by a selection procedure that delivery in specific departments, such as emergency and assessed the organisational support, commitment for surgery departments [8,10]. But, to our knowledge, it is participation, availability of personnel, time to realise unknown whether the QIC method itself is applicable for improvements, and experience with improvement pro- implementing complex process redesigns, which aim to jects. Each group of eight hospitals joined the pro- change patterns of interaction between departments in gramme for two years (2004 to 2006, 2005 to 2007, or order to achieve speedy and effective care from a 2006 to 2008) and participated in several QICs on dif- patient’s perspective [11]. Therefore, we explored in this ferent topics (e.g., pressure ulcers, process redesign) [12]. study whether the QIC method was applied to complex The process redesign collaborative evaluated in this process redesign projects in a process redesign collabora- study represented the third group of eight hospitals. The tive in the Netherlands. overall aim of this collaborative was to reduce the time between the first visit to the outpatients clinic and the Methods start of treatment and/or to reduce the length of in-hos- The collaborative described in this paper was part of the pital stay by 30% for selected patient groups [13]. Eigh- Dutch national quality improvement programme ‘Sneller teen project teams from the eight participating hospitals Vos et al. Implementation Science 2010, 5:19 Page 4 of 11 http://www.implementationscience.com/content/5/1/19 joined this collaborative, which started in October 2006. moments or handovers in a care process so that fewer Seventeen of these teams agreed to participate in our health care workers are involved in the process, and that independent evaluation. The enrolment of project teams each worker is involved only once per iteration of a within the evaluated QIC differed per hospital. Project process. teams took part on their own initiative or were enrolled The change agent also provided a website enabling by the hospital board, but always in agreement with the project teams to share information. Although it is external change agent. recommended for QICs to test a big change idea in one series of testing cycles [8], the external change agent Process redesign collaborative advised splitting up every planned change into smaller The evaluated collaborative used a step-by-step guide, ones that could be tested instantaneously in a series of which included the model for improvement (see Figure testing cycles based on their experiences of other colla- 3). This step-by-step guide was provided by the external boratives ( Figure 1, part B). By doing so, the external change agent. Next to this, the external change agent change agent tried to ensure that teams spent their organised five collaborative meetings to inform teams initial resources on testing changes instead of dealing about the step-by-step guide as well as about changes with barriers and resistance to change. that have worked at other sites. The presented evidence for improvement focused mainly at the introduction of Data collection a one-stop-shop, in which various visits per patient To explore the applicability of the QIC method, we eval- (diagnostic examinations, consultations, and preopera- uated the process redesign collaborative in a multiple tive screening) are planned for a single day, with the case study design [14] using complementary qualitative aim of reducing the throughput time of the diagnostic data collection methods. trajectory. Examples of other process redesign change We analysed the process redesign team education ideas that were provided are: the standardisation of care manual to learn more about the provided change ideas processes in order to reduce variation, the reduction of and step-by-step guide. Further, we held a survey the number of unnecessary steps in care processes (do among hospital staff members who took part in the pro- not provide care for which there is no evidence of effi- ject implementations (project staff members) (n = 17) cacy), the reduction of the number of planning and among project leaders (n = 17) to gather data on Figure 3 Step-by-step guide used in the process redesign collaborative including the model for improvement * The provided outcome measures were: 1) access time to outpatients clinic, 2) duration of diagnostic trajectory, 3) time between diagnosis and treatment, and 4) length of in-hospital stay. The provided intermediate measure (an indicator of progress [21]) was the number of visits to the outpatients clinic up to the start of treatment. Vos et al. Implementation Science 2010, 5:19 Page 5 of 11 http://www.implementationscience.com/content/5/1/19 project characteristics and aims, composition of the pro- All project teams intended to make improvements in ject teams, and project plans (including (planned) waiting times and delays, but in different areas (access changes, project progress, and the application of the times, throughput times of diagnostic trajectories, and/ model for improvement). The surveys also included or length of stay) and for different types of patient questions about team organisation (including a clear groups. The median value of the volume of the involved task division, self responsibility for progress, good com- patient groups was 150 patients a year (range 17 to pliance to arrangements, good communication and 1,000). The number of medical departments involved in coordination, be in charge of implementation), organisa- the redesigned care process was on average three and tional support (including support of strategic manage- varied per project from one to eight departments. In ment, organisational willingness to change) and external seven instances, not all medical departments involved change agent support (including sufficient support and participated in the project team. supply of instruments, transfer of valuable insights), because it is known from literature that these are pre- Presence of preconditions for successful use of conditions for successful use of the QIC method the QIC method [12,15,16]. In the survey among project leaders, we The project leaders and project staff members of six included a validated questionnaire to assess these three project teams shared the opinion that preconditions for preconditions [15]. Project staff members were asked to successful use of the QIC method–i.e., ‘team organisa- rate the amount of organisational support and external tion’, ‘organisational support’,and ‘external change change agent support on a scale of 0 to 10. Question- agent support’–were sufficiently present (project no. 1, naires were sent to respondents one year after the start 4, 6, 10, 16, and 17). The remaining project teams show of the collaborative (September 2007), and sixteen pro- a diverse picture of the presence of the preconditions. ject staff members (response = 94%) and eleven project In general, almost all project teams were positive about leaders (response = 65%) completed and returned them. the organisation of their project team. One-half of the We also interviewed all project staff members (n = 17) project teams had the opinion that support from their after they returned the questionnaire between October organisation and/or external change agent support was and December 2007. Interview themes were: change lacking. agent support (provided best practices, change concepts, and quality improvement methods), shared experiences Evaluation of the collaborative process between teams, and applicability of the model for This section describes the collaborative process accord- improvement. ing to the step-by-step guide provided to the process In addition, we observed the guidance and training redesign collaborative (see figure 3). offered by the external change agent during meetings and training sessions of the process redesign collabora- Step one tive. The observations provided us context for the analy- All projects started with a process analysis of the exist- sis of the questionnaires and interviews. ing care process. Sixteen of the seventeen projects per- Finally, we analysed the results reported on the out- formed a baseline measurement. come and intermediate measures set by the external change agent, who collected these results in a ‘Sneller Step two Beter’ database and, at our request, provided us with The baseline measurement and ideas about the desired these data (December 2007). care process formed the input for the project aims and All gathered information was used to describe the col- changes that needed to be implemented. Although all laborative process and to assess the applicability of the project teams formulated project aims, only fourteen QIC method to process redesign. Additional information formulated at least one specific and measurable aim about the preconditions was gathered to evaluate (range 0 to 7, average 2) (see Table 2). whether they could have influenced the results. Step three Results After setting aims, the next step was to establish mea- Characteristics of the process redesign projects sures that would indicate whether a change led to an within the collaborative improvement. With one exception, all project teams Table 1 gives an overview of the characteristics of the made use of one or more of the outcome measures pro- process redesign projects. Fifteen project teams chose to vided for the effect measurement. The provided inter- redesign an elective care process. Eight of those projects mediate measure was used by eleven project teams involved care for cancer patients. Two project teams (Table 2). For three teams, this measure (number of vis- chose to redesign an acute care process. its to outpatient clinic) was not applicable because these Vos et al. Implementation Science 2010, 5:19 Page 6 of 11 http://www.implementationscience.com/content/5/1/19 Table 1 Characteristics of enrolled process redesign projects 1,2 Volume of patient Acute (A) or Process to be redesigned Involved medical departments 1 1,2 1,2,3 group (patients/yr) elective (E) (description*, N**) No. Patient Group Access Diagnostic In- to care trajectory hospital (outpatients stay clinic) 1. Acute stomach 200 A - - + Internal medicine; Radiology; 3 (2) complaints Pathology 2. Breast cancer 120 E - + - Oncology; Surgery; Radiology 3 (2) 3. Breast cancer 250 E + + - Oncology; Surgery; Radiology 3 (?) 4. Chronic ?E - + + Lung diseases 1 (1) Obstructive Pulmonary Disease 5. Colon cancer 110 E + + + Gastroenterology; Surgery; 5 (4) Oncology; Anaesthesiology; Radiology 6. Colon cancer 80 E + + - Gastroenterology; Surgery; 4 (?) Radiology; Pathology 7. Colon cancer 150 E - + - Gastroenterology; Surgery; 5 (2) Radiology; Anaesthesiology; Oncology 8. Head- and neck 650 E + + + Ear, Nose and Throat; Radiology; 8 (5) cancer Jaw surgery; Radiotherapy; Oncology; Pathology; Anaesthesiology; Plastic Surgery 9. Hematuria 130 E + + + Urology; Radiology 2 (2) 10. Lung cancer 400 E - + - Lung diseases; Radiology; Surgery; 5 (1) Pathology; Anaesthesiology 11. Oesophageal 17 A - - + Paediatric Surgery; Intensivist; 3 (2) atresia (children) Radiology 12. Open Chest 1000 E + - + Thorax Surgery; Anaesthesiology 2 (2) Surgery 13. Small 250 E + + - Orthopaedics; Radiology 2 (2) Orthopaedic interventions 14. Small >200 E + + - Orthopaedics; Anaesthesiology 2 (1) Orthopaedic interventions 15. Benign Prostate 100 E - + + Urology 1 (1) Hypertrophy 16. Colon cancer 100 E + + + Surgery; Gastroenterology; 4 (1) Radiology; Oncology 17. Varicose veins 150 E + + - Surgery; Dermatology 2 (2) + Yes, - No; * in bold: medical departments that are represented by a medical specialist in the project team; ** number of medical departments involved 1 2 (number of medical departments represented in project team). Data source: interviews among project staff members. Data source: survey among project staff members. Data source: survey among project leaders. projects involved only the redesign of in-hospital stay. for instance: Is the date of surgery planned directly after For two project teams, the provided intermediate mea- setting the diagnosis, yes or no? Five projects used no sure was not applicable because it was not related to the additional intermediate or process measure at all. Rea- project aims: namely, the project did not strive to sons for not using project-specific measures were that reduce the number of visits. teams thought the provided measures gave enough Eight project teams established additional outcome insight to know whether a change is an improvement or measures: for example, time between several diagnostic because their project aims were not considered measur- examinations within the diagnostic trajectory. Six project able (e.g., qualitative aims such as a standardised dis- teams appointed intermediate and/or process measures charge planning, or appointing one contact person for to establish whether a process change was accomplished, the patient during the whole care process). Vos et al. Implementation Science 2010, 5:19 Page 7 of 11 http://www.implementationscience.com/content/5/1/19 Table 2 Application of the model for improvement in the enrolled process redesign projects 1 1,2,3 1,3 Key elements Specific and Measures of improvement Key changes PDSA Effect of the model measurable measurement for aims (N) (collaborative improvement goals reached?)? Provided by Established by the Evidence for Supplied external change project team improvement (one- change agent stop-shop) concepts implemented in used? redesign? No. Patient Outcome Inter- Outcome Process Yes/ Comments Group mediate and/or No intermediate 1. Acute + (1) + n.a. + + - n.a. + - - (?) stomach complaints 2. Breast cancer - (0) + + - - - Already + - - (?) implemented 3. Breast cancer + (1) - - - - - Already + . - (?) implemented 4. Chronic + (1) + - - + - One-stop-shop + + - (?) Obstructive is no solution Pulmonary for the existing Disease bottleneck 5. Colon cancer + (4) + + + - + - + - - (?) 6. Colon cancer + (1) + + - - + - + - - (?) 7. Colon cancer - (0) + + + - + - + +/-* - (?) 8. Head- and + (7) + + + + + - + - - (?) neck cancer 9. Hematuria + (2) + + - - + - + . - (?) 10. Lung cancer + (2) + - + - +/- Three-stop-shop + + - (?) 11. Oesophageal - (0) + n.a. - - - n.a. + - - (?) atresia (children) 12. Open Chest + (6) + n.a. + + - n.a. + - - (?) Surgery 13. Small + (2) + + + + + - + + - (?) Orthopaedic interventions 14. Small + (3) + + - + + - + + - (?) Orthopaedic interventions 15. Benign + (2) + + - - + - + - + (+) Prostate Hypertrophy 16. Colon cancer + (5) + + + - +/- Three-stop-shop + - + (+) 17. Varicose veins + (5) + + - - + - + - + (+) 1 2 3 4 Data source: survey among project staff members. Data source: survey among project leaders. Data source: interviews among project staff members. Data source: Sneller Beter database. + Yes, - No, . missing data, n.a. non applicable, because project only involves in-hospital care. * This project team used PDSA for testing and implementing a selection of the changes. Step four they already combined all visits in the diagnostic trajec- The main change idea, the one-stop-shop, presented in tory into one; they did not redesign a diagnostic trajec- the collaborative meetings was applicable for 11 project tory at the outpatients’ clinic; or the long throughput teams (Table 2). Two of them did not succeed in com- time wasnot aresultofmanyvisitsbut of alongwait- bining the visits in one day due to organisational charac- ing list for one specific diagnostic examination. All pro- teristics, the nature of the needed diagnostics, and/or ject teams applied one or more of the other provided the burden of the diagnostics to the patients. Six project change concepts to redesign their care processes. Appli- teams thought the evidence was not applicable because cation of these change ideas required that project teams Vos et al. Implementation Science 2010, 5:19 Page 8 of 11 http://www.implementationscience.com/content/5/1/19 first investigated the causes of waiting times and delays cycle, hadnot yetmeasuredany interimresults by in the redesigned process and then tailored the change December 2007 (one year after the start of the QIC). ideas to their own setting. However, according to the Therefore it is unknown whether they reached the colla- project staff, tailoring change ideas proved more difficult borative goals. in care processes in which more medical departments From this description of the collaborative process we were involved, and accordingly more disagreement can identify several difficulties experienced by the pro- existed between the involved medical departments about ject teams in applying the QIC method to process rede- the changes that had to be made. sign. First, the adoption of change ideas and the accompanying measures provided by the external Steps five and six change agent, appeared not (directly) applicable for During the interviews, project staff members were asked these collaborative project teams. Project teams had to whether they had applied the PDSA cycle for change. tailor change ideas to their own context or could not Five confirmed that their project team used or was use the provided change ideas at all. going to use the PDSA cycle. However, these five project Second, the adoption of the model for improvement teams did not split up every planned change in smaller by the project teams was hampered. Project teams were changes as the change agent suggested. Further, staff not capable of testing change ideas within a short time members of these five project teams indicated that the frame using PDSA cycles due to: the need for tailoring PDSA cycle was not or would not be performed in a change ideas to their own context, and the complexity rapid cyclical mode because both the preparation for the of aligning several interests of involved medical depart- test as well as the test of the change itself was time con- ments; the small volumes of the involved patient groups; suming. Because the patient groups were relatively and hospital information systems that proved unable to small, a testing cycle took considerable time even when generate data on the appointed measures. the number of patients per testing period was scaled Third, project teams did not experience peer stimulus. down. The use of the PDSA cycle was also hampered by All collaborative project teams intended to make the fact that hospital information systems proved unable improvements on an administrative subject, but in dif- to generate data on the appointed measures when more ferent parts of care processes (access times, throughput hospital departments were involved. As a consequence, times of diagnostic trajectories, and/or length of stay) project teams had to gather data by hand, which was for different types of patient groups. As a consequence, time consuming. project teams saw few similarities between their projects, Theteamsthatdid notuse or were notgoing to use rarely shared experiences, and demonstrated no compe- PDSA for implementation (n = 10) chose to change the titive behaviour. organisation of the care process radically by implement- Further, a number of project teams perceived a lack of ing their ‘newly designed process’ at once without first organisational support and external change agent sup- testing the individual changes. According to these project port. However, the project teams that succeeded in teams, testing change ideas within a short timeframe was implementing changes (projects 15, 16, and 17) shared not applicable to their situation because of the number of the opinion that preconditions for successful use of the medical departments involved and/or the small number QIC method–i.e., ‘team organisation’, ‘organisational of patients involved in their redesign. Another reason for support’,and ‘external change agent support’–were in nottesting in rapidcycles wasthe feelingthatatest general sufficiently present. Only organisational support could fail due to non-optimal conditions when support- lacked in one of the three project teams (project 15). ing processes were not optimised. For example, the team implementing changes in the care for open chest surgery Discussion patients considered it impossible to test a new operating From the results it seems that in the evaluated colla- room planning process. Changing the planning system borative the QIC method was not used. Apparently, it for the operating room would necessitate adjusting all did not contribute to empower project teams to imple- the supporting processes, including the working hours of ment their process redesign in a short timeframe. As a the teams and how the rooms were prepared. Any testing consequence, this study could not show whether the before the altering of supporting processes would be QIC method can effectively contribute to process rede- massively disruptive. sign, if used. The description of the collaborative pro- cess provides us with valuable information about the Step seven difficulties experienced by the project teams in applying Three project teams performed an effect measurement the QIC method to process redesign. In this section, we and reached collaborative goals (Table 2). The other will discuss explanations for these difficulties, which project teams, including those that used the PDSA concentrate on a lack of fit between the QIC method Vos et al. Implementation Science 2010, 5:19 Page 9 of 11 http://www.implementationscience.com/content/5/1/19 and process redesign, a non-optimal application of the In this study, the complexity of aligning department QIC method, and non-optimal conditions for using the schedules and interests became more apparent when the QIC method. number of departments involved in a care process increased. The project teams might have improved the Non-optimal fit between the QIC method and process collaboration across boundaries if they had included in redesign their team a medical specialist from all medical depart- First, a lot of the project teams needed customised solu- ment(s) involved. However, the need for buy-in solutions tions for their process redesign, while the QIC method before testing a change could also be due to the fact that aims to spread standardised evidence-based practices or the external change agent advised splitting up every change ideas to serve many teams at the same time with planned change into smaller changes. Although smaller a limited number of external change agents. According changes can reduce the risk of failure, it also lowers the to the QIC method collaborative project teams should expectations of the benefits of a change. Unclear or smal- benefit of the exchange of the standardised change ideas ler benefits do not stimulate medical departments to in such a way that they can eliminate much of the invest in making changes. investigative work on problem analysis and change ideas Difficulties in using the PDSA cycle meant that most in comparison with traditional quality project teams [3]. teams decided to implement changes without testing For example, in a QIC for pressure ulcers, an external them. Subsequently, teams did not get feedback on the change agent can provide concrete best practices from work they were doing and did not experience a momen- pressure ulcer guidelines to perfect the elements of care, tum of change [18]. It is known from previous studies such as ‘minimise skin pressure through the use of a that consistent ongoing measurement is required to tell positioning schedule for clients with an identified risk whether changes being made are leading to an improve- for pressure ulcer development’. This best practice can ment, and to provide basis for continued action [19,20]. then be tested and, if it works, be implemented directly Because of this lack of feedback, teams were not stimu- in every setting. Process redesign, however, calls for cus- lated to adapt another change idea for improvement, tomised solutions because project teams need to handle whichinturnsloweddownthe implementation of context-specific causes of waiting times and delays in changes. care processes determined by the existing interaction Although the difficulties with the use of the PDSA patterns between departments in their hospital. Project cycle are (almost) inevitable in process redesign projects teams can therefore not test the standard change ideas in functionally organised hospitals, the use of the PDSA provided by the change agent within a short time frame could beimproved bytakingcareof some precondi- but have to investigate the causes of waiting times and tions. First, hospital information systems should be able delays and to tailor change ideas to their own setting. to generate data on the appointed measures. Second, the As a consequence, the collaborative cannot eliminate number of patients involved in the care process that the investigative work on problem analysis and profit need to be redesigned has to be big enough to test a from standard change ideas provided by the external change idea within a number of days. change agent as the QIC method prescribes. Second, the model for improvement, and especially the Non-optimal application of the QIC method PDSA cycle, seemed inappropriate to test intended Next to the non-optimal fit between the QIC method changes within a short timeframe. The QIC method and process redesign, difficulties can also be due to the assumes that testing one big change idea lowers the resis- selection process of the collaborative project teams. The tance to a change because clinicians are more likely to be external change agent included project teams in the col- reassured that the change is effective [8,17]. This assump- laborative that worked on different parts of care pro- tion ignores the fact that testing changes that affect sev- cesses (access times, throughput times of diagnostic eral departments may lead to more consultation before trajectories, and/or length of stay) for different types of testing a change and thus to an increased possibility of patient groups, while the QIC method aims to imple- resistance to a change. This happened in the hospitals ment evidence-based practice through sharing knowl- involved as result of their functional structure, in which edge with others in a similar setting [4]. Probably, the every department has its own responsibilities and tries to external change agent could have provided peer stimu- optimise its own functioning. These functional bound- lus if it had selected project teams that worked on com- aries hampered, for example, the adjustment of the parable process redesign projects with comparable goals. department schedules needed to realise a ‘one-stop- Nevertheless, lack of peer stimulus can also occur shop’. After all, more relationships are affected, and more between comparable redesign projects because of the different interests play a role. As a result, project teams existence of context-specific causes of delays and wait- could only start testing after a buy-in or political solution. ing times. Vos et al. Implementation Science 2010, 5:19 Page 10 of 11 http://www.implementationscience.com/content/5/1/19 Non-optimal conditions for using the QIC method research into the applicability of the QIC method for Next to hospital information systems to generate data process redesign is needed. on outcome, intermediate and process measures, com- plex process redesign projects need support to change Acknowledgements interaction patterns between involved departments. A This research is funded by ZonMw, the Netherlands Organisation of Health number of project teams perceived a lack of organisa- Research and Development. tional and external change agent support, despite the Author details facts that all project teams received external change 1 NIVEL, Netherlands Institute for Health Services Research, P.O. Box 1568, agent support and the participating hospitals were 3500 BN Utrecht, the Netherlands. Impact, Dutch knowledge and advice center for post-disaster psychosocial care, P.O. Box 78, 1110 AB Diemen, the enrolled in the ‘Sneller Beter’ programme by a selection Netherlands. Institute for Health and Care Research, Department of Public procedure that assessed the organisational support. and Occupational Health, VU University Medical Centre Amsterdam, P.O. Box Unfortunately, we could not identify factors that con- 7057, 1007 MB Amsterdam, the Netherlands. Care and Public Health Research Institute (CAPHRI), Maastricht University Medical Centre+, P.O. Box tributed to this perceived lack of organisational and 5800, 6202 AZ Maastricht, the Netherlands. external change agent support. Authors’ contributions LV was responsible for designing the study, conducting the multiple case Limitations study, analyzing and interpreting the data, and drafting the manuscript. MD This study aimed to assess the applicability of the QIC participated in the design of the study, assisted in interpreting the results, method for process redesign. Although we think the and drafting the manuscript. CW and GM participated in the design of the study, assisted in interpreting the results, the critical revision of the findings of this study provide useful information for manuscript, and its supervision. All authors have read and approved the final future collaboratives, the results need to be interpreted manuscript. with caution. The findings of this evaluation could be Competing interests influenced negatively by the selection process of both The authors declare that they have no competing interests. the collaborative project teams and the care processes to be redesigned. For instance, not all teams participated in Received: 24 April 2009 Accepted: 25 February 2010 Published: 25 February 2010 the collaborative on a voluntary basis. Unfortunately, we could not determine with certainty to which project References teams this applied and how this influenced the colla- 1. Baker GR: Collaborating for improvement: The institute for Healthcare borative process. Improvement’s Breakthrough Series. New Medicine 1997, 1:5-8. 2. Lindenauer PK: Effects of quality improvement collaboratives. BMJ 2008, Another limitation is that the gathered data are not 336:1448-1449. complete. However, observations during meetings and 3. Ovretveit J, Bate P, Cleary P, Cretin S, Gustafson D, McInnes K, McLeod H, training sessions of the process redesign collaborative Molfenter T, Plsek P, Robert G, Shortell S, Wilson T: Quality collaboratives: lessons from research. Qual Saf Health Care 2002, 11:345-351. showed us that the missing data of project leaders and 4. Newton PJ, Halcomb EJ, Davidson PM, Denniss AR: Barriers and facilitators project staff members are not related to poor perform- to the implementation of the collaborative method: reflections from a ing project teams and/or organizational support. The single site. Qual Saf Health Care 2007, 16:409-414. 5. Schouten LM, Hulscher ME, van Everdingen JJ, Huijsman R, Grol RP: poor availability of effect measurements on collaborative Evidence for the impact of quality improvement collaboratives: goals can be contributed to the fact that it is not feasible systematic review. BMJ 2008, 336:1491-1494. for many project teams to redesign, implement, and per- 6. IHI: The Breakthrough Series: IHI’s Collaborative Model for Achieving Breakthrough Improvement. IHI Innovation Series white paper Boston: Institute form an effect measurement within a year, and to the for Healthcare Improvement 2003. non-optimal fit between the principles of the used QIC 7. Langley GJ, Nolan KM, Nolan TW, Norman CL, Provost LP: The Improvement method and process redesign. Guide: A Practical Approach to Enhancing Organizational Performance San Francisco: Jossey-Bass 1996. 8. Nolan TW, Schall MW, Berwick DM, Roessner J: Reducing Delays and Waiting Conclusion Times throughout the Healthcare System Boston: Institute for Healthcare This study showed that the need for tailoring standard Improvement 1996. 9. Mittman BS: Creating the Evidence Base for Quality Improvement change ideas to the context of collaborative project Collaboratives. Ann Intern Med 2004, 140:897-901. teams, and the complexity of aligning several interests 10. Kerr D, Bevan H, Gowland B, Penny J, Berwick D: Redesigning cancer care. of involved medical departments, hampered the use of BMJ 2002, 324:164-166. 11. Locock L: Health care redesign: meaning, origins and application. Qual the QIC method for process redesign. We cannot deter- Saf Health Care 2003, 12:53-57. mine whether the QIC method is appropriate for pro- 12. Dückers MLA, Spreeuwenberg P, Wagner C, Groenewegen PP: Exploring cess redesign. As result of the selection process for the black box of quality improvement collaboratives: modelling relations between conditions, applied changes and outcomes. Implement Sci 2009, participation of project teams by the external change 4:74. agent peer stimulus was non-optimal. Further project 13. Rouppe van der Voort M, Stoffer M, Zuiderent-Jerak T, Janssen S, Berg M: teams felt that preconditions for successful use of the Breakthrough Process Redesign III: 2006-2007 Better Faster pillar 3, T2S2 en T3S1 (in Dutch) Utrecht/Rotterdam/Utrecht: Quality Institute for Health Care QIC method were lacking. Therefore, additional Vos et al. Implementation Science 2010, 5:19 Page 11 of 11 http://www.implementationscience.com/content/5/1/19 CBO, Institute of Health policy and Management of the Erasmus University Rotterdam, Order of Medical Specialists 2006. 14. Yin RK: Case study research: design and methods Thousand Oaks, CA (etc.): Sage, 3 2003. 15. Dückers ML, Wagner C, Groenewegen PP: Developing and testing an instrument to measure the presence of conditions for successful implementation of quality improvement collaboratives. BMC Health Serv Res 2008, 8:172. 16. Dückers MLA, Wagner C, Groenewegen PP: Conditions for a sector wide, knowledge based, improvement programme in the Dutch hospital care (in Dutch). Acta Hospitalia 2005, 45:37-54. 17. Resar R: Why we need to learn standardisation. Aust Fam Physician 2005, 34:67-68. 18. Plsek PE: Quality improvement methods in clinical medicine. Pediatrics 1999, 103(1 Suppl E):203-214. 19. Berwick DM: Developing and testing changes in delivery of care. Ann Intern Med 1998, 128:651-656. 20. Leape LL, Rogers G, Hanna D, Griswold P, Federico F, Fenn CA, Bates DW, Kirle L, Clarridge BR: Developing and implementing new safe practices: voluntary adoption through statewide collaboratives. Qual Saf Health Care 2006, 15:289-295. 21. Gaucher EJ, Coffey RJ: Breakthrough Performance: Accelerating the Transformation of Health Care Organizations San Francisco: Jossey-Bass 2000. doi:10.1186/1748-5908-5-19 Cite this article as: Vos et al.: Applying the quality improvement collaborative method to process redesign: a multiple case study. Implementation Science 2010 5:19. Submit your next manuscript to BioMed Central and 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 redistribution Submit your manuscript at www.biomedcentral.com/submit http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Implementation Science Springer Journals

Applying the quality improvement collaborative method to process redesign: a multiple case study

Download PDF
11 pages

Loading next page...
 
/lp/springer-journals/applying-the-quality-improvement-collaborative-method-to-process-LBE6JHiuum

References (26)

Publisher
Springer Journals
Copyright
Copyright © 2010 by Vos et al; licensee BioMed Central Ltd.
Subject
Medicine & Public Health; Health Promotion and Disease Prevention; Health Administration; Health Informatics; Public Health
eISSN
1748-5908
DOI
10.1186/1748-5908-5-19
pmid
20184762
Publisher site
See Article on Publisher Site

Abstract

Background: Despite the widespread use of quality improvement collaboratives (QICs), evidence underlying this method is limited. A QIC is a method for testing and implementing evidence-based changes quickly across organisations. To extend the knowledge about conditions under which QICs can be used, we explored in this study the applicability of the QIC method for process redesign. Methods: We evaluated a Dutch process redesign collaborative of seventeen project teams using a multiple case study design. The goals of this collaborative were to reduce the time between the first visit to the outpatient’s clinic and the start of treatment and to reduce the in-hospital length of stay by 30% for involved patient groups. Data were gathered using qualitative methods, such as document analysis, questionnaires, semi-structured interviews and participation in collaborative meetings. Results: Application of the QIC method to process redesign proved to be difficult. First, project teams did not use the provided standard change ideas, because of their need for customised solutions that fitted with context- specific causes of waiting times and delays. Second, project teams were not capable of testing change ideas within short time frames due to: the need for tailoring changes ideas and the complexity of aligning interests of involved departments; small volumes of involved patient groups; and inadequate information and communication technology (ICT) support. Third, project teams did not experience peer stimulus because they saw few similarities between their projects, rarely shared experiences, and did not demonstrate competitive behaviour. Besides, a number of project teams reported that organisational and external change agent support was limited. Conclusions: This study showed that the perceived need for tailoring standard change ideas to local contexts and the complexity of aligning interests of involved departments hampered the use of the QIC method for process redesign. We cannot determine whether the QIC method would have been appropriate for process redesign. Peer stimulus was non-optimal as a result of the selection process for participation of project teams by the external change agent. In conclusion, project teams felt that necessary preconditions for successful use of the QIC method were lacking. Background based practice through sharing knowledge with others Quality improvement collaboratives (QICs) are used in a similar setting over a short period of time [4]. increasingly in many countries to achieve large-scale Within the QIC method, external change agents provide improvements in performance and to provide specific collaborative project teams from different healthcare remedies to overcome the typically slow diffusion of departments or organisations with a clear vision for medical and healthcare innovations [1-3]. A QIC is a ideal care in the topic area and a set of specific changes multifaceted method that seeks to implement evidence- that may improve system performance significantly [5,6]. Project teams also learn from the external change agent about the model for improvement. The model for * Correspondence: [email protected] improvement incorporates four key elements [6]: speci- NIVEL, Netherlands Institute for Health Services Research, P.O. Box 1568, fic and measurable aims; measures of improvement that 3500 BN Utrecht, the Netherlands © 2010 Vos et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. A P S D A P S D A P S D A P S D A P A P S D A P S D Vos et al. Implementation Science 2010, 5:19 Page 2 of 11 http://www.implementationscience.com/content/5/1/19 are tracked over time; key changes that will result in the In addition to the relatively efficient use of external desired improvement; and series of parallel testing plan- change agent support and the exchange of change ideas do-study-act (PDSA) cycles. Each series involves a test as well as the model for improvement, the strength of of onechangeidea(Figure1, part A) [7]. On thebasis the QIC method seems to be that collaborative project of the results of the first test of one series, a project teams share experiences of making changes, which team can decide to refine the change idea (in case the accelerates the rate of improvement (peer stimulus) [3]. change idea works in their context) or to start a new However, despite the widespread use of QICs, a recent test series of a new change idea (in case the test did not review on their impact indicates that evidence is positive lead to the desired result). These PDSA cycles should be but limited, and the effects cannot be predicted with cer- short but significant, testing a big change idea in a short tainty [5]. This apparent inconsistency requires a deeper timeframe so that a team can identify ways to improve understanding of how and why QICs work. Therefore it or change theidea[8].InFigure2,anexampleisgiven is necessary to explore the ‘black box’ of the intervention to illustrate the model for improvement. and to study the determinants of success or failure of the A. Testing and implementing changes according to the QIC methodology Example of a series of linked testing cycles Modify the protocol and Reengineered system make it standard practice Use the entire protocol with all patients T4 Modify the protocol and use it with other patients T3 Use part of a protocol with small group of T2 patients and refine it T1 Idea 1: Concept Design: Idea 2 Idea 3 Idea 4 one-stop-shop Source original figure: Langley GJ et al . (1996) [5] B. Testing and implementing changes according to the advised method in the evaluated collaborative Part a of Part b of Part c of Part d of Idea 1: One-stop-shop protocol protocol protocol protocol (e.g. triage) (e.g. planning all diagnostics in one day) Part a , Part b , Part c, Part d , Idea 2 Idea 2 Idea 2 Idea 2 Idea 2 Tekst Part a , Part b , Part c, Part d , Idea 3 Idea 3 Idea 3 Idea 3 Idea 3 Part a , Part b , Part c, Part d , Idea 4 Idea 4 Idea 4 Idea 4 Idea 4 Figure 1 Testing and implementing changes using PDSA cycles A P S D esig Detail D n A P A S D S D A P S D A P S D A P S D Concept design Reengineered system A P S D A P S D Vos et al. Implementation Science 2010, 5:19 Page 3 of 11 http://www.implementationscience.com/content/5/1/19 1. Aim --------------------------------------------------- Figure 2 Applying the model for improvement, an example QIC method [5,9]. In this article, we contribute to this by Beter’ (‘Better Faster’), which began in 2004 as an initia- assessing the applicability of this quality improvement tive from the Ministry of Health and the Dutch Hospital method to process redesign. Process redesign aims to Association. ‘Sneller Better’ aimed to realise substantial improve the organisation of care delivery in terms of and appealing performance improvements in three waiting times in a patients’ care trajectory. From other groups of eight Dutch hospitals in the areas of patient studies it is already known that the QIC method can be logistics and safety. These twenty-four hospitals were successfully applied to improve the organisation of care enrolled in the programme by a selection procedure that delivery in specific departments, such as emergency and assessed the organisational support, commitment for surgery departments [8,10]. But, to our knowledge, it is participation, availability of personnel, time to realise unknown whether the QIC method itself is applicable for improvements, and experience with improvement pro- implementing complex process redesigns, which aim to jects. Each group of eight hospitals joined the pro- change patterns of interaction between departments in gramme for two years (2004 to 2006, 2005 to 2007, or order to achieve speedy and effective care from a 2006 to 2008) and participated in several QICs on dif- patient’s perspective [11]. Therefore, we explored in this ferent topics (e.g., pressure ulcers, process redesign) [12]. study whether the QIC method was applied to complex The process redesign collaborative evaluated in this process redesign projects in a process redesign collabora- study represented the third group of eight hospitals. The tive in the Netherlands. overall aim of this collaborative was to reduce the time between the first visit to the outpatients clinic and the Methods start of treatment and/or to reduce the length of in-hos- The collaborative described in this paper was part of the pital stay by 30% for selected patient groups [13]. Eigh- Dutch national quality improvement programme ‘Sneller teen project teams from the eight participating hospitals Vos et al. Implementation Science 2010, 5:19 Page 4 of 11 http://www.implementationscience.com/content/5/1/19 joined this collaborative, which started in October 2006. moments or handovers in a care process so that fewer Seventeen of these teams agreed to participate in our health care workers are involved in the process, and that independent evaluation. The enrolment of project teams each worker is involved only once per iteration of a within the evaluated QIC differed per hospital. Project process. teams took part on their own initiative or were enrolled The change agent also provided a website enabling by the hospital board, but always in agreement with the project teams to share information. Although it is external change agent. recommended for QICs to test a big change idea in one series of testing cycles [8], the external change agent Process redesign collaborative advised splitting up every planned change into smaller The evaluated collaborative used a step-by-step guide, ones that could be tested instantaneously in a series of which included the model for improvement (see Figure testing cycles based on their experiences of other colla- 3). This step-by-step guide was provided by the external boratives ( Figure 1, part B). By doing so, the external change agent. Next to this, the external change agent change agent tried to ensure that teams spent their organised five collaborative meetings to inform teams initial resources on testing changes instead of dealing about the step-by-step guide as well as about changes with barriers and resistance to change. that have worked at other sites. The presented evidence for improvement focused mainly at the introduction of Data collection a one-stop-shop, in which various visits per patient To explore the applicability of the QIC method, we eval- (diagnostic examinations, consultations, and preopera- uated the process redesign collaborative in a multiple tive screening) are planned for a single day, with the case study design [14] using complementary qualitative aim of reducing the throughput time of the diagnostic data collection methods. trajectory. Examples of other process redesign change We analysed the process redesign team education ideas that were provided are: the standardisation of care manual to learn more about the provided change ideas processes in order to reduce variation, the reduction of and step-by-step guide. Further, we held a survey the number of unnecessary steps in care processes (do among hospital staff members who took part in the pro- not provide care for which there is no evidence of effi- ject implementations (project staff members) (n = 17) cacy), the reduction of the number of planning and among project leaders (n = 17) to gather data on Figure 3 Step-by-step guide used in the process redesign collaborative including the model for improvement * The provided outcome measures were: 1) access time to outpatients clinic, 2) duration of diagnostic trajectory, 3) time between diagnosis and treatment, and 4) length of in-hospital stay. The provided intermediate measure (an indicator of progress [21]) was the number of visits to the outpatients clinic up to the start of treatment. Vos et al. Implementation Science 2010, 5:19 Page 5 of 11 http://www.implementationscience.com/content/5/1/19 project characteristics and aims, composition of the pro- All project teams intended to make improvements in ject teams, and project plans (including (planned) waiting times and delays, but in different areas (access changes, project progress, and the application of the times, throughput times of diagnostic trajectories, and/ model for improvement). The surveys also included or length of stay) and for different types of patient questions about team organisation (including a clear groups. The median value of the volume of the involved task division, self responsibility for progress, good com- patient groups was 150 patients a year (range 17 to pliance to arrangements, good communication and 1,000). The number of medical departments involved in coordination, be in charge of implementation), organisa- the redesigned care process was on average three and tional support (including support of strategic manage- varied per project from one to eight departments. In ment, organisational willingness to change) and external seven instances, not all medical departments involved change agent support (including sufficient support and participated in the project team. supply of instruments, transfer of valuable insights), because it is known from literature that these are pre- Presence of preconditions for successful use of conditions for successful use of the QIC method the QIC method [12,15,16]. In the survey among project leaders, we The project leaders and project staff members of six included a validated questionnaire to assess these three project teams shared the opinion that preconditions for preconditions [15]. Project staff members were asked to successful use of the QIC method–i.e., ‘team organisa- rate the amount of organisational support and external tion’, ‘organisational support’,and ‘external change change agent support on a scale of 0 to 10. Question- agent support’–were sufficiently present (project no. 1, naires were sent to respondents one year after the start 4, 6, 10, 16, and 17). The remaining project teams show of the collaborative (September 2007), and sixteen pro- a diverse picture of the presence of the preconditions. ject staff members (response = 94%) and eleven project In general, almost all project teams were positive about leaders (response = 65%) completed and returned them. the organisation of their project team. One-half of the We also interviewed all project staff members (n = 17) project teams had the opinion that support from their after they returned the questionnaire between October organisation and/or external change agent support was and December 2007. Interview themes were: change lacking. agent support (provided best practices, change concepts, and quality improvement methods), shared experiences Evaluation of the collaborative process between teams, and applicability of the model for This section describes the collaborative process accord- improvement. ing to the step-by-step guide provided to the process In addition, we observed the guidance and training redesign collaborative (see figure 3). offered by the external change agent during meetings and training sessions of the process redesign collabora- Step one tive. The observations provided us context for the analy- All projects started with a process analysis of the exist- sis of the questionnaires and interviews. ing care process. Sixteen of the seventeen projects per- Finally, we analysed the results reported on the out- formed a baseline measurement. come and intermediate measures set by the external change agent, who collected these results in a ‘Sneller Step two Beter’ database and, at our request, provided us with The baseline measurement and ideas about the desired these data (December 2007). care process formed the input for the project aims and All gathered information was used to describe the col- changes that needed to be implemented. Although all laborative process and to assess the applicability of the project teams formulated project aims, only fourteen QIC method to process redesign. Additional information formulated at least one specific and measurable aim about the preconditions was gathered to evaluate (range 0 to 7, average 2) (see Table 2). whether they could have influenced the results. Step three Results After setting aims, the next step was to establish mea- Characteristics of the process redesign projects sures that would indicate whether a change led to an within the collaborative improvement. With one exception, all project teams Table 1 gives an overview of the characteristics of the made use of one or more of the outcome measures pro- process redesign projects. Fifteen project teams chose to vided for the effect measurement. The provided inter- redesign an elective care process. Eight of those projects mediate measure was used by eleven project teams involved care for cancer patients. Two project teams (Table 2). For three teams, this measure (number of vis- chose to redesign an acute care process. its to outpatient clinic) was not applicable because these Vos et al. Implementation Science 2010, 5:19 Page 6 of 11 http://www.implementationscience.com/content/5/1/19 Table 1 Characteristics of enrolled process redesign projects 1,2 Volume of patient Acute (A) or Process to be redesigned Involved medical departments 1 1,2 1,2,3 group (patients/yr) elective (E) (description*, N**) No. Patient Group Access Diagnostic In- to care trajectory hospital (outpatients stay clinic) 1. Acute stomach 200 A - - + Internal medicine; Radiology; 3 (2) complaints Pathology 2. Breast cancer 120 E - + - Oncology; Surgery; Radiology 3 (2) 3. Breast cancer 250 E + + - Oncology; Surgery; Radiology 3 (?) 4. Chronic ?E - + + Lung diseases 1 (1) Obstructive Pulmonary Disease 5. Colon cancer 110 E + + + Gastroenterology; Surgery; 5 (4) Oncology; Anaesthesiology; Radiology 6. Colon cancer 80 E + + - Gastroenterology; Surgery; 4 (?) Radiology; Pathology 7. Colon cancer 150 E - + - Gastroenterology; Surgery; 5 (2) Radiology; Anaesthesiology; Oncology 8. Head- and neck 650 E + + + Ear, Nose and Throat; Radiology; 8 (5) cancer Jaw surgery; Radiotherapy; Oncology; Pathology; Anaesthesiology; Plastic Surgery 9. Hematuria 130 E + + + Urology; Radiology 2 (2) 10. Lung cancer 400 E - + - Lung diseases; Radiology; Surgery; 5 (1) Pathology; Anaesthesiology 11. Oesophageal 17 A - - + Paediatric Surgery; Intensivist; 3 (2) atresia (children) Radiology 12. Open Chest 1000 E + - + Thorax Surgery; Anaesthesiology 2 (2) Surgery 13. Small 250 E + + - Orthopaedics; Radiology 2 (2) Orthopaedic interventions 14. Small >200 E + + - Orthopaedics; Anaesthesiology 2 (1) Orthopaedic interventions 15. Benign Prostate 100 E - + + Urology 1 (1) Hypertrophy 16. Colon cancer 100 E + + + Surgery; Gastroenterology; 4 (1) Radiology; Oncology 17. Varicose veins 150 E + + - Surgery; Dermatology 2 (2) + Yes, - No; * in bold: medical departments that are represented by a medical specialist in the project team; ** number of medical departments involved 1 2 (number of medical departments represented in project team). Data source: interviews among project staff members. Data source: survey among project staff members. Data source: survey among project leaders. projects involved only the redesign of in-hospital stay. for instance: Is the date of surgery planned directly after For two project teams, the provided intermediate mea- setting the diagnosis, yes or no? Five projects used no sure was not applicable because it was not related to the additional intermediate or process measure at all. Rea- project aims: namely, the project did not strive to sons for not using project-specific measures were that reduce the number of visits. teams thought the provided measures gave enough Eight project teams established additional outcome insight to know whether a change is an improvement or measures: for example, time between several diagnostic because their project aims were not considered measur- examinations within the diagnostic trajectory. Six project able (e.g., qualitative aims such as a standardised dis- teams appointed intermediate and/or process measures charge planning, or appointing one contact person for to establish whether a process change was accomplished, the patient during the whole care process). Vos et al. Implementation Science 2010, 5:19 Page 7 of 11 http://www.implementationscience.com/content/5/1/19 Table 2 Application of the model for improvement in the enrolled process redesign projects 1 1,2,3 1,3 Key elements Specific and Measures of improvement Key changes PDSA Effect of the model measurable measurement for aims (N) (collaborative improvement goals reached?)? Provided by Established by the Evidence for Supplied external change project team improvement (one- change agent stop-shop) concepts implemented in used? redesign? No. Patient Outcome Inter- Outcome Process Yes/ Comments Group mediate and/or No intermediate 1. Acute + (1) + n.a. + + - n.a. + - - (?) stomach complaints 2. Breast cancer - (0) + + - - - Already + - - (?) implemented 3. Breast cancer + (1) - - - - - Already + . - (?) implemented 4. Chronic + (1) + - - + - One-stop-shop + + - (?) Obstructive is no solution Pulmonary for the existing Disease bottleneck 5. Colon cancer + (4) + + + - + - + - - (?) 6. Colon cancer + (1) + + - - + - + - - (?) 7. Colon cancer - (0) + + + - + - + +/-* - (?) 8. Head- and + (7) + + + + + - + - - (?) neck cancer 9. Hematuria + (2) + + - - + - + . - (?) 10. Lung cancer + (2) + - + - +/- Three-stop-shop + + - (?) 11. Oesophageal - (0) + n.a. - - - n.a. + - - (?) atresia (children) 12. Open Chest + (6) + n.a. + + - n.a. + - - (?) Surgery 13. Small + (2) + + + + + - + + - (?) Orthopaedic interventions 14. Small + (3) + + - + + - + + - (?) Orthopaedic interventions 15. Benign + (2) + + - - + - + - + (+) Prostate Hypertrophy 16. Colon cancer + (5) + + + - +/- Three-stop-shop + - + (+) 17. Varicose veins + (5) + + - - + - + - + (+) 1 2 3 4 Data source: survey among project staff members. Data source: survey among project leaders. Data source: interviews among project staff members. Data source: Sneller Beter database. + Yes, - No, . missing data, n.a. non applicable, because project only involves in-hospital care. * This project team used PDSA for testing and implementing a selection of the changes. Step four they already combined all visits in the diagnostic trajec- The main change idea, the one-stop-shop, presented in tory into one; they did not redesign a diagnostic trajec- the collaborative meetings was applicable for 11 project tory at the outpatients’ clinic; or the long throughput teams (Table 2). Two of them did not succeed in com- time wasnot aresultofmanyvisitsbut of alongwait- bining the visits in one day due to organisational charac- ing list for one specific diagnostic examination. All pro- teristics, the nature of the needed diagnostics, and/or ject teams applied one or more of the other provided the burden of the diagnostics to the patients. Six project change concepts to redesign their care processes. Appli- teams thought the evidence was not applicable because cation of these change ideas required that project teams Vos et al. Implementation Science 2010, 5:19 Page 8 of 11 http://www.implementationscience.com/content/5/1/19 first investigated the causes of waiting times and delays cycle, hadnot yetmeasuredany interimresults by in the redesigned process and then tailored the change December 2007 (one year after the start of the QIC). ideas to their own setting. However, according to the Therefore it is unknown whether they reached the colla- project staff, tailoring change ideas proved more difficult borative goals. in care processes in which more medical departments From this description of the collaborative process we were involved, and accordingly more disagreement can identify several difficulties experienced by the pro- existed between the involved medical departments about ject teams in applying the QIC method to process rede- the changes that had to be made. sign. First, the adoption of change ideas and the accompanying measures provided by the external Steps five and six change agent, appeared not (directly) applicable for During the interviews, project staff members were asked these collaborative project teams. Project teams had to whether they had applied the PDSA cycle for change. tailor change ideas to their own context or could not Five confirmed that their project team used or was use the provided change ideas at all. going to use the PDSA cycle. However, these five project Second, the adoption of the model for improvement teams did not split up every planned change in smaller by the project teams was hampered. Project teams were changes as the change agent suggested. Further, staff not capable of testing change ideas within a short time members of these five project teams indicated that the frame using PDSA cycles due to: the need for tailoring PDSA cycle was not or would not be performed in a change ideas to their own context, and the complexity rapid cyclical mode because both the preparation for the of aligning several interests of involved medical depart- test as well as the test of the change itself was time con- ments; the small volumes of the involved patient groups; suming. Because the patient groups were relatively and hospital information systems that proved unable to small, a testing cycle took considerable time even when generate data on the appointed measures. the number of patients per testing period was scaled Third, project teams did not experience peer stimulus. down. The use of the PDSA cycle was also hampered by All collaborative project teams intended to make the fact that hospital information systems proved unable improvements on an administrative subject, but in dif- to generate data on the appointed measures when more ferent parts of care processes (access times, throughput hospital departments were involved. As a consequence, times of diagnostic trajectories, and/or length of stay) project teams had to gather data by hand, which was for different types of patient groups. As a consequence, time consuming. project teams saw few similarities between their projects, Theteamsthatdid notuse or were notgoing to use rarely shared experiences, and demonstrated no compe- PDSA for implementation (n = 10) chose to change the titive behaviour. organisation of the care process radically by implement- Further, a number of project teams perceived a lack of ing their ‘newly designed process’ at once without first organisational support and external change agent sup- testing the individual changes. According to these project port. However, the project teams that succeeded in teams, testing change ideas within a short timeframe was implementing changes (projects 15, 16, and 17) shared not applicable to their situation because of the number of the opinion that preconditions for successful use of the medical departments involved and/or the small number QIC method–i.e., ‘team organisation’, ‘organisational of patients involved in their redesign. Another reason for support’,and ‘external change agent support’–were in nottesting in rapidcycles wasthe feelingthatatest general sufficiently present. Only organisational support could fail due to non-optimal conditions when support- lacked in one of the three project teams (project 15). ing processes were not optimised. For example, the team implementing changes in the care for open chest surgery Discussion patients considered it impossible to test a new operating From the results it seems that in the evaluated colla- room planning process. Changing the planning system borative the QIC method was not used. Apparently, it for the operating room would necessitate adjusting all did not contribute to empower project teams to imple- the supporting processes, including the working hours of ment their process redesign in a short timeframe. As a the teams and how the rooms were prepared. Any testing consequence, this study could not show whether the before the altering of supporting processes would be QIC method can effectively contribute to process rede- massively disruptive. sign, if used. The description of the collaborative pro- cess provides us with valuable information about the Step seven difficulties experienced by the project teams in applying Three project teams performed an effect measurement the QIC method to process redesign. In this section, we and reached collaborative goals (Table 2). The other will discuss explanations for these difficulties, which project teams, including those that used the PDSA concentrate on a lack of fit between the QIC method Vos et al. Implementation Science 2010, 5:19 Page 9 of 11 http://www.implementationscience.com/content/5/1/19 and process redesign, a non-optimal application of the In this study, the complexity of aligning department QIC method, and non-optimal conditions for using the schedules and interests became more apparent when the QIC method. number of departments involved in a care process increased. The project teams might have improved the Non-optimal fit between the QIC method and process collaboration across boundaries if they had included in redesign their team a medical specialist from all medical depart- First, a lot of the project teams needed customised solu- ment(s) involved. However, the need for buy-in solutions tions for their process redesign, while the QIC method before testing a change could also be due to the fact that aims to spread standardised evidence-based practices or the external change agent advised splitting up every change ideas to serve many teams at the same time with planned change into smaller changes. Although smaller a limited number of external change agents. According changes can reduce the risk of failure, it also lowers the to the QIC method collaborative project teams should expectations of the benefits of a change. Unclear or smal- benefit of the exchange of the standardised change ideas ler benefits do not stimulate medical departments to in such a way that they can eliminate much of the invest in making changes. investigative work on problem analysis and change ideas Difficulties in using the PDSA cycle meant that most in comparison with traditional quality project teams [3]. teams decided to implement changes without testing For example, in a QIC for pressure ulcers, an external them. Subsequently, teams did not get feedback on the change agent can provide concrete best practices from work they were doing and did not experience a momen- pressure ulcer guidelines to perfect the elements of care, tum of change [18]. It is known from previous studies such as ‘minimise skin pressure through the use of a that consistent ongoing measurement is required to tell positioning schedule for clients with an identified risk whether changes being made are leading to an improve- for pressure ulcer development’. This best practice can ment, and to provide basis for continued action [19,20]. then be tested and, if it works, be implemented directly Because of this lack of feedback, teams were not stimu- in every setting. Process redesign, however, calls for cus- lated to adapt another change idea for improvement, tomised solutions because project teams need to handle whichinturnsloweddownthe implementation of context-specific causes of waiting times and delays in changes. care processes determined by the existing interaction Although the difficulties with the use of the PDSA patterns between departments in their hospital. Project cycle are (almost) inevitable in process redesign projects teams can therefore not test the standard change ideas in functionally organised hospitals, the use of the PDSA provided by the change agent within a short time frame could beimproved bytakingcareof some precondi- but have to investigate the causes of waiting times and tions. First, hospital information systems should be able delays and to tailor change ideas to their own setting. to generate data on the appointed measures. Second, the As a consequence, the collaborative cannot eliminate number of patients involved in the care process that the investigative work on problem analysis and profit need to be redesigned has to be big enough to test a from standard change ideas provided by the external change idea within a number of days. change agent as the QIC method prescribes. Second, the model for improvement, and especially the Non-optimal application of the QIC method PDSA cycle, seemed inappropriate to test intended Next to the non-optimal fit between the QIC method changes within a short timeframe. The QIC method and process redesign, difficulties can also be due to the assumes that testing one big change idea lowers the resis- selection process of the collaborative project teams. The tance to a change because clinicians are more likely to be external change agent included project teams in the col- reassured that the change is effective [8,17]. This assump- laborative that worked on different parts of care pro- tion ignores the fact that testing changes that affect sev- cesses (access times, throughput times of diagnostic eral departments may lead to more consultation before trajectories, and/or length of stay) for different types of testing a change and thus to an increased possibility of patient groups, while the QIC method aims to imple- resistance to a change. This happened in the hospitals ment evidence-based practice through sharing knowl- involved as result of their functional structure, in which edge with others in a similar setting [4]. Probably, the every department has its own responsibilities and tries to external change agent could have provided peer stimu- optimise its own functioning. These functional bound- lus if it had selected project teams that worked on com- aries hampered, for example, the adjustment of the parable process redesign projects with comparable goals. department schedules needed to realise a ‘one-stop- Nevertheless, lack of peer stimulus can also occur shop’. After all, more relationships are affected, and more between comparable redesign projects because of the different interests play a role. As a result, project teams existence of context-specific causes of delays and wait- could only start testing after a buy-in or political solution. ing times. Vos et al. Implementation Science 2010, 5:19 Page 10 of 11 http://www.implementationscience.com/content/5/1/19 Non-optimal conditions for using the QIC method research into the applicability of the QIC method for Next to hospital information systems to generate data process redesign is needed. on outcome, intermediate and process measures, com- plex process redesign projects need support to change Acknowledgements interaction patterns between involved departments. A This research is funded by ZonMw, the Netherlands Organisation of Health number of project teams perceived a lack of organisa- Research and Development. tional and external change agent support, despite the Author details facts that all project teams received external change 1 NIVEL, Netherlands Institute for Health Services Research, P.O. Box 1568, agent support and the participating hospitals were 3500 BN Utrecht, the Netherlands. Impact, Dutch knowledge and advice center for post-disaster psychosocial care, P.O. Box 78, 1110 AB Diemen, the enrolled in the ‘Sneller Beter’ programme by a selection Netherlands. Institute for Health and Care Research, Department of Public procedure that assessed the organisational support. and Occupational Health, VU University Medical Centre Amsterdam, P.O. Box Unfortunately, we could not identify factors that con- 7057, 1007 MB Amsterdam, the Netherlands. Care and Public Health Research Institute (CAPHRI), Maastricht University Medical Centre+, P.O. Box tributed to this perceived lack of organisational and 5800, 6202 AZ Maastricht, the Netherlands. external change agent support. Authors’ contributions LV was responsible for designing the study, conducting the multiple case Limitations study, analyzing and interpreting the data, and drafting the manuscript. MD This study aimed to assess the applicability of the QIC participated in the design of the study, assisted in interpreting the results, method for process redesign. Although we think the and drafting the manuscript. CW and GM participated in the design of the study, assisted in interpreting the results, the critical revision of the findings of this study provide useful information for manuscript, and its supervision. All authors have read and approved the final future collaboratives, the results need to be interpreted manuscript. with caution. The findings of this evaluation could be Competing interests influenced negatively by the selection process of both The authors declare that they have no competing interests. the collaborative project teams and the care processes to be redesigned. For instance, not all teams participated in Received: 24 April 2009 Accepted: 25 February 2010 Published: 25 February 2010 the collaborative on a voluntary basis. Unfortunately, we could not determine with certainty to which project References teams this applied and how this influenced the colla- 1. Baker GR: Collaborating for improvement: The institute for Healthcare borative process. Improvement’s Breakthrough Series. New Medicine 1997, 1:5-8. 2. Lindenauer PK: Effects of quality improvement collaboratives. BMJ 2008, Another limitation is that the gathered data are not 336:1448-1449. complete. However, observations during meetings and 3. Ovretveit J, Bate P, Cleary P, Cretin S, Gustafson D, McInnes K, McLeod H, training sessions of the process redesign collaborative Molfenter T, Plsek P, Robert G, Shortell S, Wilson T: Quality collaboratives: lessons from research. Qual Saf Health Care 2002, 11:345-351. showed us that the missing data of project leaders and 4. Newton PJ, Halcomb EJ, Davidson PM, Denniss AR: Barriers and facilitators project staff members are not related to poor perform- to the implementation of the collaborative method: reflections from a ing project teams and/or organizational support. The single site. Qual Saf Health Care 2007, 16:409-414. 5. Schouten LM, Hulscher ME, van Everdingen JJ, Huijsman R, Grol RP: poor availability of effect measurements on collaborative Evidence for the impact of quality improvement collaboratives: goals can be contributed to the fact that it is not feasible systematic review. BMJ 2008, 336:1491-1494. for many project teams to redesign, implement, and per- 6. IHI: The Breakthrough Series: IHI’s Collaborative Model for Achieving Breakthrough Improvement. IHI Innovation Series white paper Boston: Institute form an effect measurement within a year, and to the for Healthcare Improvement 2003. non-optimal fit between the principles of the used QIC 7. Langley GJ, Nolan KM, Nolan TW, Norman CL, Provost LP: The Improvement method and process redesign. Guide: A Practical Approach to Enhancing Organizational Performance San Francisco: Jossey-Bass 1996. 8. Nolan TW, Schall MW, Berwick DM, Roessner J: Reducing Delays and Waiting Conclusion Times throughout the Healthcare System Boston: Institute for Healthcare This study showed that the need for tailoring standard Improvement 1996. 9. Mittman BS: Creating the Evidence Base for Quality Improvement change ideas to the context of collaborative project Collaboratives. Ann Intern Med 2004, 140:897-901. teams, and the complexity of aligning several interests 10. Kerr D, Bevan H, Gowland B, Penny J, Berwick D: Redesigning cancer care. of involved medical departments, hampered the use of BMJ 2002, 324:164-166. 11. Locock L: Health care redesign: meaning, origins and application. Qual the QIC method for process redesign. We cannot deter- Saf Health Care 2003, 12:53-57. mine whether the QIC method is appropriate for pro- 12. Dückers MLA, Spreeuwenberg P, Wagner C, Groenewegen PP: Exploring cess redesign. As result of the selection process for the black box of quality improvement collaboratives: modelling relations between conditions, applied changes and outcomes. Implement Sci 2009, participation of project teams by the external change 4:74. agent peer stimulus was non-optimal. Further project 13. Rouppe van der Voort M, Stoffer M, Zuiderent-Jerak T, Janssen S, Berg M: teams felt that preconditions for successful use of the Breakthrough Process Redesign III: 2006-2007 Better Faster pillar 3, T2S2 en T3S1 (in Dutch) Utrecht/Rotterdam/Utrecht: Quality Institute for Health Care QIC method were lacking. Therefore, additional Vos et al. Implementation Science 2010, 5:19 Page 11 of 11 http://www.implementationscience.com/content/5/1/19 CBO, Institute of Health policy and Management of the Erasmus University Rotterdam, Order of Medical Specialists 2006. 14. Yin RK: Case study research: design and methods Thousand Oaks, CA (etc.): Sage, 3 2003. 15. Dückers ML, Wagner C, Groenewegen PP: Developing and testing an instrument to measure the presence of conditions for successful implementation of quality improvement collaboratives. BMC Health Serv Res 2008, 8:172. 16. Dückers MLA, Wagner C, Groenewegen PP: Conditions for a sector wide, knowledge based, improvement programme in the Dutch hospital care (in Dutch). Acta Hospitalia 2005, 45:37-54. 17. Resar R: Why we need to learn standardisation. Aust Fam Physician 2005, 34:67-68. 18. Plsek PE: Quality improvement methods in clinical medicine. Pediatrics 1999, 103(1 Suppl E):203-214. 19. Berwick DM: Developing and testing changes in delivery of care. Ann Intern Med 1998, 128:651-656. 20. Leape LL, Rogers G, Hanna D, Griswold P, Federico F, Fenn CA, Bates DW, Kirle L, Clarridge BR: Developing and implementing new safe practices: voluntary adoption through statewide collaboratives. Qual Saf Health Care 2006, 15:289-295. 21. Gaucher EJ, Coffey RJ: Breakthrough Performance: Accelerating the Transformation of Health Care Organizations San Francisco: Jossey-Bass 2000. doi:10.1186/1748-5908-5-19 Cite this article as: Vos et al.: Applying the quality improvement collaborative method to process redesign: a multiple case study. Implementation Science 2010 5:19. Submit your next manuscript to BioMed Central and 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 redistribution Submit your manuscript at www.biomedcentral.com/submit

Journal

Implementation Science – Springer Journals

Published: Feb 25, 2010

There are no references for this article.