Managerial strategies to make incentives meaningful and motivating

Current trends in health policy

The policy application of market logics in healthcare is no longer predominantly a US phenomenon but is now widely applied in publically funded European healthcare systems as well (Roland and Rosen, 2011). This is certainly the case in Sweden, where this study is set (Anell, 2015; Harrison and Calltorp, 2000). Inspired by classic economic theory, competition between providers is encouraged by allowing patients to choose (Appleby and Dixon, 2004). Furthermore, financial-incentive models are developed to increase provider performance (Saltman, 2002). However, the empirical evidence is mixed and fragmented. Increased patient choice has been associated with increased provider performance in competitive markets (Cooper et al., 2011), yet literature reviews provide limited support for stating that choice in itself improves efficiency or quality of care (Pollock et al., 2011; Fotaki et al., 2008). Similarly, review studies evaluating the impact of various financial-incentive models in healthcare conclude that the support for their effectiveness is limited (Flodgren et al., 2011; Chaix-Couturier et al., 2000). Incentives linked to specific provider behaviors, such as pay-for-performance-models, have shown to increase, e.g., productivity and cost efficiency (Eijkenaar et al., 2013). However, concurrent reports of unintended consequences for patients (Chaix-Couturier et al., 2000; Eijkenaar et al., 2013) and healthcare professionals (Campbell et al., 2007, 2008; Swarna Nantha, 2013; McDonald et al., 2007) show negative side effects. Quality-based reimbursement (Conrad and Perry, 2009) and bundled reimbursement (Mechanic, 2011) have been proposed to avoid the pitfalls of narrowly defined performance targets. This aims to empower clinicians to drive quality improvement, but evaluations show both opportunities and challenges when it is implemented in practice (Mechanic, 2011). Overall, it is difficult to draw conclusions concerning the general effects of financial incentives due to methodological weaknesses of evaluation studies (Eijkenaar et al., 2013; Flodgren et al., 2011; Chaix-Couturier et al., 2000).

In summary, whereas research exists regarding how the application of incentive models can be improved (Porter and Lee, 2013; Conrad and Perry, 2009), organizational characteristics, such as management structure, leadership and culture, are seldom considered (Frolich et al., 2007). Indeed, there is a need for conceptual development and empirical data clarifying how externally imposed incentives are filtered through organizations and translated into motivation and behavioral changes at the individual level. In particular, a need exists for theory-based models that explore the hybrid role of managers in bridging the macro and micro levels of healthcare systems who align multiple, potentially conflicting sources of staff motivation. In the following section, we will draw on theories from the fields of psychology, sociology, behavioral economics and management to discuss the literature on individual motivation and the role of managers.

Internal and external sources of motivation

Motivation is a multidimensional phenomenon defined as the energy and intention behind an action (Ryan and Deci, 2000). In contrast to classic economic theory underlying modern policy applications, motivation theories from different domains all take a broader approach to incentives. Both internal and external sources of motivation are acknowledged, and the interplay and potential conflicts between them are highlighted (Ryan and Deci, 2000; Swick, 2000; Ellingsen and Johannesson, 2007, 2008). Moreover, there is an agreement across fields that internal sources of motivation (that is, doing work for work’s sake) have a greater impact on behavior over time compared to external rewards (Deci et al., 1999; Gneezy et al., 2011; Ellingsen and Johannesson, 2007, 2008). There is also a convergence of motivational theories in that they often include basic human needs in terms of autonomy, expertise and pro-social behavior. In psychology, this is described in self-determination theory (Deci and Ryan, 2000). In sociology, theories of professionalism describe similar motivational forces that are highly applicable to healthcare. Professionals (an occupational group characterized by certain preferences) are described as being motivated by a high level of expertise (Freidson, 2001) and acting autonomously based on that expertise (Bøgh-Andersen and Holm-Pedersen, 2012). In addition, pro-social behaviors are also seen as a main component, as professionals are internally motivated to safeguard patient needs and social justice (Swick, 2000). Similar assumptions about motivational sources are reported in behavioral economics (Ellingsen and Johannesson, 2007).

The interrelationship between multiple sources of motivation

Researchers from several fields have proposed that the coexistence of multiple sources of motivation can have a negative impact on overall motivation. The introduction of monetary rewards might harm or “crowd out” internal motivation and reduce overall performance (Ryan and Deci, 2000). Similarly, pro-social behavior has been shown to decrease in response to monetary rewards (Gneezy et al., 2011; Benabou and Tirole, 2003). Some have argued that the risk of crowding-out effects is higher in public service such as healthcare due to the pro-social preference of staff (Frey et al., 2013). There is empirical evidence to support the existence of a special kind of public service motivation (Bright, 2008) characterized by a stronger preference for internal reward, as compared to the private sector (Crewson, 1997; Georgellis et al., 2011).

In contrast, there is also empirical evidence that external rewards can be highly motivating and “crowd in” overall motivation (Frey, 1994; Frey and Jegen, 2001). According to self-determination theory, the motivational effect of external rewards is dependent upon the extent to which it is internalized and consistent with the individual’s values (Deci and Ryan, 2000). Studies evaluating different forms of rewards show that the best motivational strategy involved combining personal monetary rewards with managerial feedback (Stajkovic and Luthans, 2003). These findings suggest that managerial behavior can play an important role in creating coherence between external rewards and internal values.

To summarize, in contrast to policy applications inspired by classic economic theory, research on motivation suggests that external incentives may reduce healthcare professionals’ motivation. However, contradictory findings indicate that external incentives may mobilize motivation when aligned with individual values. The presence or absence of alignment hence seems to be a key issue. In the next section, we discuss the role of managers as intermediaries potentially linking policy to the motivational preferences of their staff.

The role of managers as intermediaries between external rewards and motivation

Management matters a great deal for organizational performance and staff motivation (Hales, 1999), including in healthcare settings (Lega et al., 2013). Similarly, studies of line managers show that they play a key role in the implementation of new care processes (Dopson and Fitzgerald, 2006). According to traditional conceptualizations, managers’ primary role is to control staff through the planning, organization and coordination of tasks (Hales, 1999). More contemporary models of managerial behavior emphasize the reciprocal nature of management. In these models, managers are described as agents and facilitators between a network of stakeholders both inside and outside of the organization (Hales, 2002).

The number of managers employed to bridge the gap between top management and professionals has increased in healthcare (Kuhlmann and von Knorring, 2014). The rise of managerialism has conventionally been regarded as conflicting logic in professional organizations (Brommels, 2010), but research shows that hybrid managerial roles held by professionals is a common phenomenon (Kulhmann et al., 2013). Hybrid managers enable professional and managerial logic to be intertwined by facilitating interaction between multiple stakeholders in organizations (Postma et al., 2015). Such integration has been shown to improve both organizational performance and quality outcomes. Thus, the mere presence of management is insufficient, but management being informed by professional knowledge and values may have positive effects (Lega et al., 2013). However, more knowledge is necessary on the specific strategies that managers use to balance various stakeholder perspectives in healthcare (Kuhlmann and von Knorring, 2014).

The central task of coordinating different stakeholder perspectives in healthcare has been described as articulation work (Strauss, 1988; Corbin and Strauss, 1993). Articulation work includes coordinating and fine tuning all tasks that emerge from the care process and comprises the communication and alignment of different stakeholder perspectives (Grant et al., 2015). Thus, it is not a separate management process but rather an integral component of healthcare management based on professional expertise and culture (Postma et al., 2015). Despite the centrality of articulation work in healthcare, its role in adapting to new policy is still largely unknown (Postma et al., 2015).

To summarize, one central managerial behavior is to align the needs of different stakeholders by capturing perspectives ranging from the macro to the micro level. Building on this, we propose that the role of managers is central to policy implementation, and their strategies merit further exploration and analysis.

Toward a tentative conceptual model

The literature on individual motivation and management from the fields of psychology, sociology and behavioral economics is summarized in Figure 1. The conceptual model in Figure 1 suggests that managers can be understood as key intermediaries who use strategies to align the interests and perspectives of different stakeholders (e.g. policy makers, the organization and staff). It focuses on the role of managers and does not cover other contextual aspects that may mediate the relationship between external rewards and the individual motivation of staff.

Setting

The study targets a patient-choice reform in specialized care for hip and knee replacement that was introduced in 2009 and is still active. The County Council of Stockholm, a politically elected regional authority with responsibility for healthcare provision (Anell, 2015) for approximately 2.2 million people, introduced the reform. Acknowledging that governance commonly refers to a broad concept of the regulatory functions of a healthcare system (Kuhlmann and Burau, 2008), this reform includes a specific governance model built on specialization of care and provider competition. The reform includes a bundled reimbursement model that allows clinical freedom in designing the episode of care. The specific governance model of this reform will henceforth be referred to as “the model.” The model applies to both hospitals and specialized private providers with no productivity limits. The main reasons behind the reform were to increase providers’ competition through patient choice and lower waiting times for surgery. Evaluations of the effects have shown increased productivity and access to care (Vårdanalys, 2014).

The model covers a well-defined episode of care in hip and knee replacements for low-risk patients only. The selection of low-risk patients is performed using the American Society of Anesthesiologists (ASA) classification guidelines. Low-risk patients (ASA 1 & 2) may be treated by their chosen provider. However, high-risk patients (ASA 3 & 4) are excluded from the model and only handled by hospitals. Providers are reimbursed with a bundled payment to cover costs for the entire episode of care, including the final assessment for surgery, brief post-operative care and follow-up. Providers are financially responsible for all complications within two years after surgery, such as reoperations and infections. However, if infections occur, care must be provided at a hospital. If a specialized provider performed the initial surgery, they will be held accountable for the cost of care the hospital provided.

Procedures

We invited all providers involved to participate in the study by contacting operation managers through e-mail and follow-up phone calls. In dialogue with operation managers at six volunteering providers, we used a purposive sampling approach to recruit respondents for interviews. In total, 17 interviews were held between June and November 2014, each lasting 45-60 minutes. In all, 16 interviews were conducted face-to-face and one over the phone. All were recorded with a digital audio recorder. All respondents were informed about the study both orally and in writing. They were told that participation was voluntary, and they all gave their written informed consent. All respondents were informed that the focus of the study was their personal experience and were encouraged to share their personal reflections. A semi-structured interview guide was developed to address three main themes: the respondents’ understanding of the model, managerial strategies to manage the implementation of the model and the respondents’ understanding of their staff’s motivation. The semi-structured interview guide included open-ended questions; examples of questions are “In what way do you adapt your leadership in relation to the model?” and “Is there anything you do to motivate your staff to engage in activities required to make the model work in practice?” Respondents were encouraged to freely expand their reasoning, as the guide did not include pre-defined follow-up questions. The interview guide was piloted in the first two interviews of the study. Based on the respondents’ feedback, the interview guide was adjusted by making questions more specific and shortening the list of pre-defined questions, leaving more room for the respondents’ own reflections. The first author conducted all interviews. The local ethical committee authorized the study (ref. no. omitted for review).

Study population

We aimed to include different provider types in the study to obtain broad knowledge about how the same model works in different provider contexts. Two specialized providers and four hospitals volunteered to participate in the study. Two specialized providers declined to participate due to extensive internal reorganization. The participating specialized providers and one hospital are for-profit organizations. The county council owns and runs the remaining three hospitals. All participating hospitals also handled hip and knee replacements for high-risk patients under a separate provider contract. In dialogue with the operation managers, we recruited healthcare professionals in managerial positions handling the episode of care covered by the model. We aimed for a broad representation of professionals and reached out to both nurses and orthopedic surgeons in managerial positions. The operations managers were also asked to participate. In total, 18 people were asked to participate, and of these, 17 agreed to participate. The respondents’ respective roles were operations manager (six), clinical manager (five), manager of post-operative care (three), operations coordinator (one), quality manager (one) and research manager (one). All respondents had substantial clinical experience as nurses (five) and orthopedic surgeons (12), and all were clinically active.

Data analysis

All interviews were transcribed verbatim and analyzed in NVivo. Data familiarization occurred concurrent with interviewing. After completing the majority of the interviews (15), data saturation was reached. The final two respondents were recruited to control for additional themes, but no new themes were discovered. After the in-depth immersion of data, we applied a two-step data-analysis process using a hypo-deductive approach (Fereday and Muir-Cochrane, 2008). First, the first author made a deductive abstraction of a focused data set. Guided by general definitions of managerial behavior (Hales, 1999), data of interest were identified throughout the data corpus. To minimize the risk of bias, the last author validated the abstracted data through independent identification of relevant data in three sampled interviews. In total, 85 percent of the identified data overlapped perfectly; all remaining inconsistencies were discussed until a consensus was reached. The second step of the analysis focused on the data set addressing managerial behavior using a thematic analysis with an inductive approach (Braun and Clarke, 2006). The remaining data corpus served as background information and provided a deeper understanding of managerial behavior. First initial codes such as “avoid communicating cost” and “talk about patient value” were generated. Thereafter, the codes were clustered to identify initial themes (e.g. “adapting,” “sense making”). Initial themes were reviewed and revised iteratively, eventually resulting in defining and naming final themes, i.e., the strategies of managers. The first author compiled the inductive thematic analysis, repeatedly discussing it with colleagues. The second and last authors validated the final themes through independent reading of three interview transcripts. The accuracy and completeness of the final themes in relation to the data were examined, and all comments were discussed until a consensus was reached.

Contextual knowledge

The respondents’ managerial strategies were based on knowledge obtained in their hybrid roles as both managers and professionals. This knowledge was a key precondition for their ability to formulate successful alignment strategies in this specific organizational context. The managers expressed a broad, experienced-based knowledge of the model’s consequences and had a clear picture of what kind of changes the model required to maintain a financially viable organization. In general, the model put pressure on providers to work more efficiently: the reimbursement levels were lower, and they were held financially accountable for any complications to provide an incentive for high-quality care. The managers also had an understanding of how the model and the changes it required affected the organization of care, the working conditions of the staff and the experiences and outcomes of the patients at their clinic. This in-depth understanding of the practical implications of the model guided managers in identifying appropriate opportunities for change and helped them formulate reasonable and realistic ambitions.

When devising strategies to achieve these ambitions, the managers’ understanding about what was needed to engage and motivate staff was key. Expressions such as “we” and “us” were used repeatedly when describing the motivational preferences of their team, indicating that staff and managers had largely the same preferences: concern for patient needs and ambition to provide high-quality care and the sensible use of resources. Professional pride in one’s expertise was also perceived as being central to staff, as was a sustainable work environment.

Managerial alignment strategies

Implications for practice

The present study has practical implications for both healthcare provision and health policy. First, healthcare providers may in different ways support managers in applying alignment strategies. In addition to time and knowledge, administrative systems with access to high-quality data on patient outcomes and resource use could increase their opportunities to give meaningful and thus motivating feedback. Providers could also maintain a dialogue with decision makers at a policy level, thus increasing opportunities for the improvement of policies. The present results further support the benefits of hybrid roles in healthcare management (Kuhlmann and von Knorring, 2014), which could guide provider organizations in their recruitment and development of management.

Second, our findings suggest that the way health policies and governance models are designed may influence managers’ abilities to form alignment strategies. Policies that support high quality of care will be easier for managers to align with staff motivation and more likely affect staff behavior and provider performance. Policy makers could also inform about new policies and governance models in a timely fashion and provide argument for the reform referring to patient and staff benefits, not only efficiency.

Our study also provides perspectives on how the process of designing policies and governance models can be improved. Healthcare is typically described as a complex adaptive system involving interaction between multiple agents and resulting in low predictability. Deviations from standardized procedures are everyday occurrences, and flexible control systems have been recommended (Sturmberg et al., 2012). It may be naïve to assume that anyone, policy makers included, can foresee all potential consequences and design optimal models. As Casalino (Casalino, 1999) noted, the inbuilt incompleteness of measures to define and control healthcare processes makes unintended consequences inevitable. The present results suggest that the introduction of new policies and governance models should be regarded as a continuous process of change rather than episodic, which requires constant adaptation to a complex environment (Weick and Quinn, 1999). In complex settings, the co-production of services involving several stakeholders sharing a common goal has been shown to be essential. This approach is increasingly applied in healthcare services, e.g., in patient-centeredness and self-management (Batalden et al., 2015). We propose that policy makers and professional managers form partnerships to co-produce policies, aligning stakeholder perspectives and enabling continuous improvement over time. To summarize, this requires a shift from efforts to design the best model to efforts to design the best process to improve the model through co-production.

Methodological considerations

The present study has limitations that require consideration. The study was conducted in specialized orthopedics, in which managers handled this specific governance model. The managerial strategies identified in this study needs to be further explored in in other healthcare settings and systems, including more complex diagnoses, different market compositions, provider structures and different incentive models. Furthermore, the body of literature guiding the focus of this study mainly takes an individual perspective on motivation and management. Particularly, system dynamics at the micro, meso or macro level needs to be investigated in future studies.

Nevertheless, our findings show similarities across groups of respondents from different provider types (private and public) who have different professional backgrounds (physicians and nurses) and roles (operations manager, clinical manager, etc.). This supports the relevance of the identified managerial strategies. To uncover the generalizability of strategies across e.g., provider types and professional groups is beyond the scope of this study but merits further exploration. In light of the expressed need for comprehensive theoretical models in health policy research (Frolich et al., 2007), we hope the tentative conceptual model presented here will be found helpful in this regard.

Regarding the validity of our results, the fact that we rely on data that only managers have provided should be considered. We have not interviewed staff or policy makers, and thus, our results reflect the views of managers rather than an objective “truth.” We have not made observations to assess managerial behaviors and thus have a limited knowledge of when and to what extent the described strategies are used. More research is needed to include complementary perspectives and additional data sources illustrating the relationships between managerial strategies and their impact on motivation.

Two operations managers from specialized providers declined to participate, and our sample includes a slightly lower proportion of respondents from specialized providers. There is a risk that our findings are colored by a selection bias in our sample of volunteering respondents. We have not been able to control to what extent our respondents have had formal training or varying skills in implementing policy reforms of this kind. There is also a risk that our respondents are more favorable or critical to policy compared to managers in general, which could affect their strategies. In summary, our empirical result highlights the need to further address the interaction between stakeholders at the policy and provider levels.