Abstract

PURPOSE We tested for direct and indirect effects that performance-based reimbursement (PBR) in primary care has on perceived individual and organizational quality of care, and the role of illegitimate tasks and moral distress as potential mediators.

METHOD We used results from the Longitudinal Occupational Health survey in Healthcare Sweden with data collected in 2021, 2022, and 2023. The sample of primary care physicians who answered at all 3 years and were aged 68 or less was 433. Performance-based reimbursement was measured using a single item. The Bern Illegitimate Tasks Scale measured illegitimate tasks, and moral distress was measured with a 10-item scale. Six items from the English National Health Staff Survey were used to measure the quality of individual and organizational care.

RESULT Of the 433 participants, 70% reported that PBR negatively impacted their work. Performance-based reimbursement was negatively associated with illegitimate tasks (b = −0.160; 95% CI, −0.240 to −0.080) and moral distress (b = −0.134; 95% CI, −0.210 to −0.058). These work factors were in turn associated with both individual and organizational quality of care. Using mediation models, we found an indirect effect (b = 0.011; 95% CI, 0.004 to 0.021) but no direct effect (b = 0.062; 95% CI, −0.019 to 0.143) between PBR on the quality of individual care.

CONCLUSION Performance-based reimbursement systems should account for the experience of individual primary care physicians to ensure effective, safe, and quality care, as this study shows how the level of illegitimate tasks and moral distress due to a PBR system can undermine care delivery. Consequently, it is imperative for stakeholders to consider how health care systems relate to the health care staff’s experience, well-being, and the care being provided.

Key words:INTRODUCTION

New public management is an approach to running public service organizations that embraces performance monitoring systems focused on quantifying outcomes.1 For health care organizations, effective and efficient quality of care is a key performance output that is regularly measured.2–4 In Sweden, the health care system is universal, publicly funded, and is organized under 21 decentralized political assemblies that set the goals for health care organizations and determine how results toward the goals are measured.5 Goals are operationalized through performance management and measurement systems, which are processes designed to assess, analyze, and improve the performance of individuals, teams, departments, or entire organizations. There is increasing criticism about the lack of effectiveness of new public management;6 however, little is known about the consequences of these systems on physicians’ experience of work and on care outcomes.

A common form of performance management and measurement systems is performance-based reimbursement (PBR).7 This is a control system in which clinics are reimbursed for services provided with the intention of providing effective and efficient care. In Sweden, PBR operates at the clinic level and is not a financial incentive system for individual physicians. Use of PBR has been found, however, to lead to shorter patient visit times and selection of low-risk or less-ill patients.8-10 This has resulted in physicians arguing against PBR as it undermines current medical standards, practices, and values.9-11

The increased use of PBR systems has been associated with increased formalization and bureaucratization,1 leading to a large amount of work for preparation and reporting of performance metrics,1,12 and physicians reporting less autonomy and poorer well-being.13 A qualitative study of Swedish physicians8 highlighted how PBR systems increased the amount of illegitimate tasks (ie, tasks beyond the scope of an employee’s primary responsibilities); unreasonable tasks (ie, tasks outside a professional’s role); and unnecessary tasks (ie, tasks not anticipated for a particular position).14,15 These tasks include completeing additional paperwork, entering specific details and diagnoses into data systems, booking separate consultations for separate conditions of one patient, and physicians seeing patients that a nurse could treat.8

In a PBR system, increases in illegitimate tasks reduce time available for clinical work8 which could create a feeling of moral distress (eg, feelings of stress or guilt due to forced unethical decisions).16,17 Although long understood in the health care sector, recognition of moral distress gained substantial traction during the COVID-19 pandemic.18,19 The focus, however, has been on the demands from trauma exposure that do not allow health care workers to be able to give the level of care that they want to.19 In this study, we apply moral distress to the context of PBR, responding to calls to examine moral distress in relation to business and corporate structures.20 Reduced time spent on core clinical tasks due to more administrative (ie, illegitimate) tasks21,22 could cause feelings of moral distress due to being unable to provide the care needed by the patients, and undermining their belief in the quality of care they are providing.23,24

While moral distress has been postulated to be linked to patient care,16,18,25 to date there has been limited empirical testing of the relationship. One cross-sectional study reported no association between moral distress and satisfaction with care quality among hospital physicians.26 The quality of care measure in this study contained items pertaining to care provided by the clinican and the health care team. As moral distress is closely linked to an individual’s appraisal of their personal situation and control of it,17 the appraisal of their own behaviors is more proximal than outcomes at the care unit level—which could be reliant on the actions of others.27,28 A negative appraisal of PBR with its increase of illegitimate tasks and feelings of moral distress could also strengthen the belief that the full potential of the clinic is not being utilized. This harms patients and health care workers, and could undermine the level of care provided at the organizational level.

Growing evidence attests to the link between poor working conditions of physicians and low standards of care being provided, which have been attributed to poor safety culture and the withdrawal of resources and attention from patient care.29–31 As the health care organizational system and structure have been positioned as an antecedent to the working conditions of the individual physicians and their well-being,28,32 it is plausible to position PBR as an antecedent to the working experience of physicians.

We used a longitudinal cohort design to test the interrelationships between PBR, quality of care, illegitimate tasks, and moral distress. This design accounts for the typical lag that the working environment has on worker health and patient care.31 Here, we distinguish between the quality of care physicians perceive they provide personally vs the quality that their organization provides.27 This is to examine if there may be a difference in how PBR systems (organizational level) and moral distress and illegitimate tasks (individual level) relate to the perception of care among different organizational levels. In this study, we test the direct and indirect effect of PBR systems on the perceived quality of care provided by the individual and the organization, where illegitimate tasks and moral distress are positioned as mediators.

METHODS

The data was pulled from the Longitudinal Occupational Health survey in Healthcare Sweden.8,33 Statistics Sweden (SCB) was responsible for sampling and data collection. From March through May 2021 (Phase 1 [baseline]), SCB sent a survey invitation to a representative sample of 6,699 physicians listed in the Swedish Occupational Registers in 2020 and 2,761 (41.2%) responded. Of all the physicians surveyed, 3,259 (48%) were registered as working in primary care and 1,013 (31.1%) of them responded. More detail of the baseline data collection is available elsewhere.33 Follow-up surveys were administered to the original cohort of responding physicians (excluding those who retired, died, or migrated during the study period) from March through May 2022 (Phase 2 [year 1]) and October through December 2023 (Phase 3 [year 2.5]). At each phase, SCB sent an initial postal invitation letter with credentials to log in to a web survey hosted by SCB. Three reminders were sent to those who had not responded. The second reminder included a paper version of the survey.

All study participants were informed that participation was voluntary and they could drop out anytime. They agreed to participate by answering and submitting the survey. The Swedish Ethical Review Authority reviewed and approved this study (Dnr: 2020 to 06613; 2021 to 05574 to 02; 2022 to 03105 to 01; 2023 to 04100 to 02).

Measurements

Figure 1 illustrates the 3 phases of the study and their and measurements. At baseline, PBR was assessed using a single item that asked respondents to rate if the PBR system impacts them positively or negatively, with answers on a 4-point scale ranging from very negative to very positive. The development of the question is described elsewhere.8 Participants who responded that they were not affected by PBR were computed as the middle value. This resulted in a 5-value item where the middle value represented those not affected by PBR (ie, neutral), a high value indicated a positive impact from PBR, and a low value indicated negative impact.

Figure 1.

Study Model Measuring the Impact of the Performance-Based Reimbursement Systems on Quality of Care

Illegitimate tasks and moral distress were measured at the Phase 1 survey. We used the 8-item Bern Illegitimate Tasks Scale15 rated on a 5-point Likert scale where a higher value indicates more illegitimate tasks and a grand mean was computed (α = 0.84). We measured moral distress using an instrument developed for Norwegian physicians,34 which we had translated into Swedish by a bilingual individual who worked in both Norwegian and Swedish health care systems. Respondents rated 10 items about how stressful each was on a 4-point Likert scale (ie, 0 for not at all; 3 for very stressful). A grand mean was computed across all 10 items where a higher value indicated more stress (α = 0.85). Earlier studies using these measures of moral distress35 and illegitimate tasks8 in the Swedish context support their validity.

In the Phase 3 survey (2.5 years after the baseline), we measured quality of care using 6 items from the English National Health Staff Survey27,36 that were translated into Swedish by a bilingual individual with knowledge of both health care contexts. These were separated into 2 dimensions of 3 items, each related to the perceived quality of care provided by the individual (eg, “I am able to deliver the patient care I aspire to”; α = 0.82) and by the organization (eg, “If a friend or relative needed treatment, I would be happy with the standard of care provided by this organization”; α = 0.71). Responses to each item were on a 5-point Likert scale (ie, 1 for strongly disagree; 5 for strongly agree) and a grand mean was calculated for each dimension where a higher score represents better perceived quality of care.

Gender and age at baseline were included as covariates after retrieval from the Longitudinal Database on Education, Income, and Occupation at Statistic Sweden. We also included quantitative demands as a covariate, given its strong influence on the constructs within the study.31,37 This was measured at baseline using 3 items (α = 0.88) from the third version of the Copenhagen Psychosocial Questionnaire.38 All reliability tests showed that the construct had a Cronbach’s α above 0.70, indicating good to high internal consistency.39

Analytical strategy

Data analysis was carried out using SPSS Statistics for Windows version 28.0 (IBM Corp). To test for serial mediation, we used Model 6 of the PROCESS macro v4.2,40 where PBR (baseline) was the predictor variable, and illegitimate work tasks and moral distress were mediators (Phase 2). Two separate models were run, 1 for each of the quality-of-care constructs (Phase 3). We used a calculation of 5,000 bias-corrected bootstrapped 95% CIs to test the indirect associations of both mediators. Participants’ age and gender, and quantitative demands were included as covariates.

RESULTS

In total, 454 primary care physicians responded at all 3 time periods. We further restricted respondents to those under the retirement age (ie, age 68 years or less), resulting in a final sample of 433 physicians. Comparisons by SCB show no systematic differences between the missing data in the sample and that of the population.

The study sample included a larger share of female (61.9%) primary care physicians than males (38.1%). At baseline, the mean age was 45 years. In the group, 36.5% had up to 10 years of experience working as physicians, 23.8% had 10 to 15 years of experience, and 39.7% had more than 15 years. Most (53.5%) responding primary care physicians reported working less than 40 hours a week, while 30.6% reported working 41 to 45 hours per week.

Of the 433 in the study sample, 1 (0.2%) reported that the PBR system very positively impacted them, while 29 (6.9%) reported a positive impact. In contrast, negative impacts were reported much more often with 249 (58.9%) responding that the PBR negatively impacted them and 52 (12.3%) reporting very negative impact. The remaining 92 (21.7%) participants reported no impact from the PBR system. Table 1 presents descriptive statistics for the measurement scores and age.

Table 1.

Descriptive Statistics for Measurements and Age

Complete serial mediation was observed between PBR and the quality of individual care provided by the individual physician, adjusting for age, gender, and quantitative workload (Table 2). Results show an indirect effect of PBR on the quality of individual care via illegitimate work tasks and moral distress (b = 0.011; 95% CI, 0.004-0.021), although there was no direct effect of PBR on the quality of individual care (b = 0.062; 95% CI, −0.019 to 0.143). PBR did negatively predict illegitimate work tasks (b = −0.160; 95% CI, −0.240 to −0.080) and moral distress (b = −0.134; 95% CI, −0.210 to −0.058). In addition, illegitimate work tasks (b = −0.229; 95% CI, −0.326 to −0.132) and moral distress (b = −0.203; 95% CI, −0.300 to −0.106) had a direct effect on the quality of individual care provided.

Table 2.

Direct and Indirect Effects Between Impact of PBR System on Quality of Care

No serial mediation was observed where the quality of organizational care was the outcome variable (b = 0.004; 95% CI, −0.001 to 0.015) adjusting for age, gender, and quantitative workload. There was no direct (b = 0.055; 95% CI, −0.039 to 0.149) effect between PBR and the quality of care provided by the organization. Unlike the quality of individual care, there was also no relationship observed between moral distress and the quality of organizational care (b = −0.109; 95% CI, −0.225 to 0.008). Table 2 shows an indirect effect between PBR and the quality of organizational care through illegitimate work tasks (b = 0.046; 95% CI, 0.019-0.077).

DISCUSSION

Providing high-quality care is a key aspect of health care services. This 3-phase longitudinal study shows that PBR negatively impacts physician’s perceived quality of care through illegitimate tasks and moral distress in Swedish primary care.

The results demonstrate the pathways by which organizational processes and systems are associated with the care being provided.

The findings raise concern that despite expected benefits (ie, increased internal and external efficiency)41,42 of new public management in public organizations in Sweden, it may be detrimental to the working experience and well-being of primary care physicians, and undermine their perception of the care provided. Although PBR systems are salient to health care provision and known to increase physicians’ administrative work,8 few studies have explored PBRs’ impact on perceived quality of care. Our findings emphasize that organizational goals must be grounded in medical practices at the operational level and metrics of success have to encompass a broader perspective that accounts for both staff experience and patient care.28,43 Quality of care must be viewed in the context of organizational demands and the conditions that primary care physicians need to carry out clinical work.

Our results should be viewed against the backdrop of an increase in unreasonable and unnecessary primary care physicians work over recent decades.44,45 This added work has the effect of not only increasing the quantitative workload, but also created increased levels of moral distress among physicians who do not feel able to provide the quality of care they should be providing. This pathway of PBR systems, illegitimate tasks, moral distress, and perceived quality of care is congruent with the theoretical46,47 and empirical31,48 literature advocating a link between organizational systems and outcomes. While this study focused on the perceived quality of care, illegitimate tasks and moral distress are central aspects of health care workers’ environment related to burnout, sickness absence, and turnover.22,49,50 Although beyond the scope of this study, it raises the question of how PBR systems may impact other measures involving the health care workforce. This may cause a vicious circle, where the PBR-related administrative work leads to poor working conditions and well-being with subsequent sickness absence and turnover, which in turn increases the administrative work of those that remain.

This study also raises the question of the suitability of current processes to monitor and manage performance, and measure quality in health care. Traditionally, health care quality has been measured by population health, care experiences, and low costs.51 This was later expanded to the Quadruple Aim, including physicians’, and other health care workers’, well-being.43,52 While these aims fill an important role in health care, they disregard the significance, impact, and state of physicians’ working conditions.27,32,43 It is thus clinically relevant to acknowledge system-level structures and processes that impact the work of primary care physicians. Future studies should continue to research how various performance management and measurement systems impact primary care physicians’ working conditions and quality of care.

Limitations

The findings of this study have several limitations. First, attrition over the study period needs to be addressed. In total, 1,510 physicians answered the survey at all 3 time points (22.5% of the baseline sample of 6,699; 54.7% of the 2,761 baseline respondents). Drop-outs were due to death, migration, retirement, parental leave, illness, or a change to employement outside of health care. In addition, by including only those who worked in primary care facilities for the full study period, we need to account for individuals who may have moved to other clinics or onto staffing agencies. It is important to note, however, that our response rates are in line with studies with similar samples,53–55 and that SCB analyzed missing data at baseline comparing the sample to the population and found no systematic differences.

Second, self-report surveys risk response and recall biases, although our use of longitudinal data at 3 phases limits the risk of common method bias. Third, primary care physicians rating individual and organizational quality care is both a strength and a limitation. Previous research has pointed out the need to encompass health care workers’ perspectives on quality care.24 Perceived quality of care may be affected, however, by an individual’s state of mind or level of stress, and the evidence of its relationship with actual patient care outcomes remains inconsistent.28 The measure of quality of care has also not been validated in the Swedish context, and studies like this are important to support its use in Sweden. Finally, PBR was measured using a single question that assesses participants’ general assessment of its impact. Using additional items that are less evaluative and more descriptive could yield a more valid representation of the construct. In addition, to obtain a PBR score, we recorded participants’ responses on 2 separate items. A clearly labeled 5-point scale where the middle number is labeled as not being affected would have improved the validity of the item and possibly reduced the positive skew within the distribution.

CONCLUSION

This longitudinal 3 phase survey of 433 physicians shows that PBR in Swedish primary care indirectly impacts perceived poor individual and organizational quality of care. The level of illegitimate tasks and moral distress due to a PBR system can undermine care delivery. The clinical implication of this study is that quality of care is not limited to best practices and evidence-based medicine, but also relies on primary care physicians’ work systems.

In practice, it would be desirable that the use of PBR accounts for the experience of individual physicians to ensure good quality care. The identification of illegitimate tasks and moral distress as potential mediators present targeted intervention opportunities to support better implementation of PBR and mitigate detrimental effects. For example, intervention studies to reduce bureaucracy and non-core tasks have been found to improve health care workers’ well-being.56 Overall, it is imperative for stakeholders across the political and health care management sphere to consider how health care systems relate to the health care workforce’s experience on the job, well-being, and the care being provided.

Footnotes

Annals Early Access article

Conflicts of interest: authors report none.

Read or post commentaries in response to this article.

Funding support: This study was supported by grants from the Swedish Research Council for Health, Working, and Welfare (#2019 to 00311) and the Swedish Research Council (#2022 to 00806).

Received for publication April 10, 2024.Revision received October 17, 2024.Accepted for publication November 19, 2024.© 2025 Annals of Family Medicine, Inc.