We conducted a cross-sectional survey in China from April to May 2024 to collect the necessary data. To understand which factors or combinations thereof are most important for perceived compensation satisfaction, we used qualitative comparative analysis (QCA). Although QCA is traditionally associated with qualitative research, its application in this study aligns with a mixed methods approach.

We designed the questionnaire according to the aim of our study. First, we conducted a pilot survey on a small sample (n=10) in Beijing, Shandong, Hainan, Jiangsu, and Ningxia that included representatives of different cultures across North, South, East, and West China, respectively. The participants in the pilot surveys were colleagues or acquaintances of the researchers. We then revised the questionnaire according to the pilot test feedback. Pilot participants suggested clarifying terms such as compensation scheme fairness, leading to revised wording in the final questionnaire (eg, I think the department’s compensation scheme is fair). Second, our sample size meets the robustness requirements for QCA [13,14]; so, the research members tried their best to recruit more physicians to complete the questionnaire before data collection was scheduled to terminate. All questions were mandatory. The survey was anonymous, and the answers were sent directly to the database to ensure confidentiality.

This study was approved by the ethics committee of China-Japan Friendship Hospital (approval 2022-KY-218) and conducted in accordance with the World Medical Association Declaration of Helsinki. Informed consent was obtained from all participants via an online consent form on the questionnaire’s opening page (participants could only proceed after clicking “agree”), and all retained the right to opt out at any stage without adverse consequences. The survey was anonymous, with all participant data fully anonymized and deidentified. Participants who completed the survey received ¥1 in electronic compensation via WeChat, equivalent to US $0.14 at the time of data collection (April-May 2024).

The questionnaire consisted of three parts: sociodemographic characteristics (eg, sex, age, place of residence, education level, technical rank, specialty, and monthly salary), compensation scheme transparency, and perceived compensation satisfaction. The physicians rated compensation transparency and perceived compensation satisfaction on a 5-point Likert scale, with strongly disagree (score 1), disagree (score 2), neither agree nor disagree (score 3), agree (score 4), and strongly agree (score 5). Lower perceived compensation satisfaction was defined as an average score of less than 3 on the compensation perception questions. Higher perceived compensation satisfaction was defined as an average score equal to or greater than 3 on the compensation perception questions [15,16]. Cronbach α and the Kaiser-Mayer-Olkin value for the questionnaire were 0.955 and 0.924, respectively, indicating high total reliability and validity. All items loaded significantly on a single factor with loadings greater than 0.6, and the cumulative variance explained was 65.3%. These results collectively support the use of these constructs as unidimensional scales.

The questionnaire collected various sociodemographic characteristics of the physician participants to provide a comprehensive background for understanding their perceived compensation satisfaction. These characteristics included sex, categorized as male and female; age, recorded in years and further classified into several ranges (≤30, 31-40, 41-50, ≥51); education level, classified as bachelor and below, masters, and doctoral degree; technical rank, grouped into resident, attending physician, and (associate) professor; department, with specialties such as surgery, internal medicine, obstetrics and gynecology, pediatrics, traditional Chinese medicine, general practice, and emergency department; monthly salary, categorized into different ranges (≤¥3000; ¥3001-¥6000; ¥6001-¥10,000; ¥10,001-¥20,000; >¥20,000); and weekly work hours, classified as the average number of hours worked per week. All variables were collected via the online questionnaire.

Compensation scheme transparency was assessed through four questions: (1) in my opinion, I am very familiar with the compensation scheme; (2) when I want to learn about the compensation scheme, I can easily find out; (3) I am able to learn about the compensation scheme in a timely manner; and (4) senior leaders proactively share the compensation distribution details with me. The Cronbach α for this dimension was 0.947, indicating relatively high reliability.

On the basis of the comparative theory, perceived compensation satisfaction was assessed from multiple perspectives: external comparison, internal comparison, input–output comparison, expectation–reality gap, and compensation scheme perception. These were measured using 7 questions: (1) compared with other professions in the hospital, my compensation level is reasonable; (2) compared with physicians in other departments, my compensation level is reasonable; (3) compared with senior physicians in my department, my compensation level is reasonable; (4) compared with physicians of the same level in my department, my compensation level is reasonable; (5) my compensation is consistent with my efforts and contributions; (6) the actual compensation is in line with my expectation; and (7) I think the department’s compensation scheme is fair. The Cronbach α for this dimension was 0.944, indicating relatively high reliability.

We used the survey platform Questionnaire Star to host our questionnaire and distributed a generated link through WeChat. Each WeChat account could respond to the questionnaire only once (to avoid repeated responses) unless modified. Our trained research members were assigned to different geographical regions, including Beijing, Hainan, Shandong, Jiangsu, and Ningxia. Each sample hospital is a tertiary public hospital and has approximately 500 to 1000 physicians. The research members visited the sample hospitals to recruit physicians face-to-face through convenience sampling until the sample size was reached. Meanwhile, they explained the purpose of the study, informed the physicians of the anonymity of the survey, and stated that it would take about 10 minutes of their time if they were available. If the physicians agreed to participate in the survey, the research members would immediately present a QR code and help the physicians scanning it. The informed consent form was also on the opening page. Only after reading and clicking the “Agree” button could participants proceed to the questionnaire. Researchers were also available to respond to questions that prospective participants had regarding unclear terms or the questionnaire items. Compensation of ¥1 (US $0.14) was offered to participants who completed the survey. We also ensured a balance of sex, age, technical rank, and specialty to ensure the representativeness. The only participation exclusion was work experience of less than 6 months.

The response rate was approximately 80.2% (401/500). In total, 401 physicians responded to the survey. Among the questionnaires received, 76 were excluded owing to unreasonable response durations or logistic errors. Ultimately, 325 questionnaires were included in our analysis. Based on QCA recommendations, 10-80 cases are required for robust analysis. Our sample size of 325 far exceeds this threshold, ensuring sufficient statistical power.

SPSS software (version 24.0; IBM Corp) was employed to assess the reliability and validity of the questionnaire, conduct factor loading analysis, and compare the differences among groups.

As perceived compensation satisfaction has several complex influences rather than a single impact, we used fuzzy-set qualitative comparative analysis (fsQCA) to explore different combinations of factors. QCA was selected to analyze how combinations of factors jointly influence satisfaction. Traditional regression models assume linear relationships, whereas QCA identifies nonlinear interactions and multiple pathways to outcomes. This approach bridges qualitative and quantitative methods [17]. We used the fsQCA 3.0 software with a package developed by Charles Ragin and Sean Davey for this analysis.

We calibrated the raw data of the selected variables following previous research [14]. First, we determined the three calibration points that represented (1) fully belong, (2) the cross-point, and (3) fully do not belong. For this, we encoded the Likert scale data of strongly agree as fully belong, neither agree nor disagree as the cross-point, and strongly disagree as fully do not belong. Following the method of preset scale anchor points as thresholds [18,19], education level was set as 3, 2, and 1, and technical rank as 4, 2.5, and 1. Continuous variables were designated for 3 qualitative breakpoints (0.95, 0.5, and 0.05), the data were converted into values between 0 and 1 [20-22]. By setting these 3 thresholds, the original data were transformed into fuzzy scores ranging from 0 to 1 by using the Calibrate (x, n1, n2, n3) function in the fsQCA software.

The initial phase of QCA involves assessing the necessity of a singular condition as well as its absence for the full adoption of the phenomenon. A condition was considered sufficient when the consistency level exceeded 0.8. Furthermore, when the consistency level exceeded 0.9, this condition was deemed essential [20-24].

The next phase of QCA is configuration analysis, which was applied to our truth table to reveal the sufficiency of the perceptions associated with different configurations of multiple conditions. To calculate the complex, parsimonious, and intermediate solutions, the consistency threshold was set to 0.8 and the case frequency threshold to 1.

Statistical significance was assessed using P values, with a threshold of P<.05. Group differences (eg, technical rank, department, monthly salary) were analyzed via 2-sided chi-square tests for categorical variables and 2-sided t tests for continuous variables.

Of the 325 participants, 163 (50.2%) were males and 162 (49.8%) were females. The majority were aged either 31-40 years (131/325, 40.31%) or 41-50 years (133/325, 40.92%). The ratio of residents to attending physicians to (associate) professors was approximately 1:2:2.3. Among the physicians, 49.8% (162/325) earned more than ¥10,000 (US $1378) per month. The average weekly work hours were 52.30 (SD 12.26) hours. Table 1 presents the sociodemographic characteristics of the participants.

Overall, most physicians felt that their compensation schemes were transparent (see Table 2). The average score for perceived compensation transparency was 3.19 (SD 1.09). Their responses indicated that they had the means to learn more about the schemes and could do so promptly. They indicated that they were familiar with their compensation schemes and that their leaders were available to provide compensation details.

As shown in Table 3, significant differences emerged in expectation–reality gap scores (2.70) compared with input–output comparisons (2.85; P<.05), suggesting that unmet compensation expectations drive dissatisfaction; the compensation gap within a department was perceived as being more reasonable than that among departments. The perceived compensation satisfaction was the lowest when compared with expectation, followed by the comparison of input and output.

We set the consistency level from 0.8 to 0.85 and 0.73 to examine the robustness of the results. Compared with those before adjustment, the configuration paths, coverage, and consistency did not change substantially after adjustment. Thus, the results are robust.

QCA enables the analysis of complex causal configurations in quantitative data, making it suitable for exploring interactions between multiple factors influencing compensation satisfaction [17].

Through the application of QCA, we analyzed how combined factors influence physicians’ perceived compensation satisfaction. We found that compensation scheme transparency was a necessary condition for higher compensation satisfaction. Physicians who were more familiar with their compensation schemes, whether actively or passively, found it easier to learn about the scheme promptly and were more satisfied with their compensation. This finding aligns with a prior study by Chinese scholars investigating rural physicians' compensation satisfaction, which found that township hospital health care workers were not satisfied with their salary unless they were aware of the salary scheme design [27]. Research conducted in the United States has found that the gender salary gap among surgeons led to burnout among female surgeons, and while compensation transparency reforms narrowed this gap, disparities persisted. Consequently, the researcher promoted a transparent, performance-based, and objective compensation plan [28,29]. In China, physician compensation comprises fixed and variable components. The fixed part, which is primarily determined by years of work experience and one’s professional title, constitutes a minor proportion of the total compensation with narrow disparities among physicians. The variable part, typically based on performance, including clinical workload, medical quality, patient satisfaction, and academic research outputs, ranges from 50% to 70% of total compensation [30,31]. Therefore, regarding compensation scheme transparency, the most important aspect is the variable part design and associated feedback. To improve physicians’ perceived compensation satisfaction and work behavior, hospital managers should maintain transparent performance assessment criteria and provide clear performance feedback.

The determinants of compensation perception were heterogeneous between junior and senior physicians. For junior physicians, if the compensation scheme is transparent, they would be satisfied even with lower pay or longer working hours. In contrast, they prioritized the transparency of the compensation scheme. For senior physicians, transparent compensation combined with a relatively high salary contributed to high compensation satisfaction. In the case of a lower salary, senior physicians’ satisfaction was promoted by a transparent compensation scheme, a higher technical title, and fewer work hours. For junior physicians, neither salary nor work hours seemed to be significantly associated with their compensation satisfaction. Therefore, junior physicians differ from other young adults at a similar career stage, who are particularly interested in higher salaries and work-life balance and unwilling to pay more than a minor amount for training to enhance their skills and knowledge [32,33]. The difference may originate from the fact that the salary of physicians increases over time with training [34]. Furthermore, the unique nature of the medical profession and the complexity of medical knowledge could explain why junior physicians do not feel the need to pursue higher salaries at this stage of their career. Moreover, junior physicians are eager to acquire more professional skills to enhance their competence in the field [35], and they understand that this initially requires long work hours and a lower salary. A survey of junior physicians in Sudan has found that their training program served as a nonmonetary incentive to trade off portions of their salaries [36]. Given the priority on professional skills, training plans offered to junior physicians should be carefully designed, and measures to address supervision-related issues are recommended in the short term from a cost perspective. For senior physicians, in addition to compensation scheme transparency, monthly salary, technical rank, and work hours all combined to affect compensation satisfaction. High entry-level requirements for physicians regarding education, training, and certification are common worldwide. Thus, we may hypothesize that high entry barriers would lead to higher salaries [37-39]. Logically, physicians should earn a relatively high compensation, as their salary is an essential component of their occupation that enables them to fulfil family responsibilities. Our findings are in line with previous research reporting that higher salaries can lead to higher satisfaction and stimulate employee enthusiasm for work [40,41]. Chinese scholars have found that low salaries relative to high training costs and low compensation for medical staff strongly incentivize overtreatment, contributing to increased medical service expenditures [42]. Given the entry-level requirements and low compensation among Chinese physicians, increasing their salaries is a significant incentive to ensure their satisfaction and avoid opportunistic behavior.

Additionally, we found that physicians with a higher technical rank and fewer work hours had higher compensation satisfaction, even when their salaries were low. This higher satisfaction with compensation may stem from Maslow’s Needs Theory, which suggests that the highest level of need is the realization of self-worth, and once the needs of a higher level are fulfilled, the more basic needs such as salary become less important [43,44]. Regarding technical rank, a survey conducted in 161 US universities also reported that associate professors were less satisfied in their jobs than full professors [45]. Thus, the title of professor serves as an approval of worth and could be the goal for many physicians. Therefore, to ensure higher compensation satisfaction, hospital managers should pay attention to the technical promotion pathway for physicians. Notably, due to the cross-sectional design, the next step will require a cohort study to validate the findings of this research.

This study has limitations. As the physicians were selected through convenience sampling, the nonrandomness of this technique could have introduced limiting factors. Although we strictly controlled sex, age, technical rank, and specialty distribution to ensure sample balance across these four key aspects, the inherent nature of convenience sampling still limits sample representativeness, and the generalizability of the conclusions requires further validation. For example, the samples were mainly from physicians willing to participate in the survey at sampled hospitals, which may lead to self-selection bias. Meanwhile, the results may be overestimated or underestimated because of the self-reporting design of our questionnaire. There are some potential influencing factors such as occupational risk that should be considered. Future research is required that employs more rigorous sampling methods, incorporates larger and more diverse samples across multiple institutions, and accounts for more potential influencing factors—all of which would enhance the generalizability of our findings and the stability of our results.

Our study considered the dynamic effects of multiple factors on physician compensation satisfaction and provided management advice to hospitals and other institutions looking to retain personnel. We found that the factors associated with compensation satisfaction differed between junior and senior physicians. For junior physicians, compensation scheme transparency, not salary level, influenced their perceived compensation satisfaction. For senior physicians, salary level was an important factor but not consistently the only decisive factor influencing satisfaction. A higher technical rank could offset the adverse effects of lower salaries. Therefore, hospital managers should focus on improving the transparency of compensation schemes and customizing benefit packages for specific employees.