Understanding medical corruption in China: a mixed-methods study

Abstract

Medical corruption is a significant obstacle to achieving health-related Sustainable Development Goals. However, the understanding of medical corruption is limited, especially in developing countries. As the largest developing country, China is also plagued by medical corruption. By employing a mixed-methods design and combining data from three resources, this study attempts to examine patterns of medical corruption in China, explore its key drivers and investigate the perceived effectiveness of recent anti-corruption interventions. Using extracted data from 3546 cases on the China Judgments Online website between 2013 and 2019, we found that bribery, embezzlement and insurance fraud accounted for 68.1%, 22.8% and 9.1% of all medical corruption cases, respectively. Bribery was the major form of medical corruption. Approximately 80% of bribe-takers were healthcare providers, and most bribe-givers were suppliers of pharmaceuticals, medical equipment and consumables. Using a nationally representative household survey, we further found that the prevalence of informal payments from patients remained at a low level between 2011 and 2018. In 2018, only 0.4% of outpatients and 1.4% of inpatients reported that they had ever given ‘red envelopes’ to physicians in the past. Finally, we conducted interviews with 17 key informants to explore drivers of medical corruption and investigated the perceived effectiveness of recent anti-corruption interventions in China. Interview results showed that financial pressure and weak oversight were two main reasons for corrupt behaviours. Interview results also suggested that the anti-corruption campaign since 2012, the national volume-based procurement, and the special campaign against medical insurance fraud had reduced opportunities for medical corruption, implying China’s positive progress in combating medical corruption. These findings hold lessons for anti-corruption interventions in China as well as other developing countries.

Medical corruption, bribery, embezzlement, insurance fraud, informal payment, health system research, mixed methods, China

Key messages

• Bribery was the major form of medical corruption in China.

• Approximately 80% of bribe-takers were healthcare providers, and most bribe-givers were suppliers of pharmaceuticals, medical equipment and consumables.

• Financial pressure and weak oversight were two main reasons for corrupt behaviours.

• The anti-corruption campaign since 2012, the national volume-based procurement, and the special campaign against medical insurance fraud were perceived to be effective in suppressing medical corruption in China.

Luming Yu, the director of the Beijing municipal health committee, is suspected of severe violations of discipline and law, and is currently undergoing disciplinary review and investigation by the Central Commission for Discipline Inspection and the National Commission of Supervision. (The official Xinhua News Agency, Beijing, 16 April 2022).

Introduction

Corruption, defined as the abuse of entrusted power for private gain, is widespread in many health systems (Transparency International, 2019). It presents in a range of ways and adversely affects health system performance. It limits people’s equal access to healthcare services and undermines the health system’s ability to provide financial protection (Gaitonde et al., 2016). Direct financial loss through medical corruption is estimated to account for at least 6% of global healthcare expenditure (García, 2019). It also has negative consequence on health outcomes, such as higher infant and children mortality, worsening antibiotic resistance and slower progress in controlling communicable and chronic disease (Mostert et al., 2015; Rönnerstrand and Lapuente, 2017; Dincer and Teoman, 2019; Farzanegan and Hofmann, 2021). Thus, medical corruption is regarded as a major threat to achieving health-related Sustainable Development Goals (Mackey et al., 2018).

To understand and assess medical corruption, existing studies have proposed several conceptual frameworks (Vian, 2008; de Sardan, 2013; Couffinhal and Frankowski, 2017). These frameworks have been applied to identify and characterize the main types of medical corruption, its key drivers and the effect of anti-corruption interventions in various countries (Huss et al., 2011; Vian et al., 2012; Rispel et al., 2016; Onwujekwe et al., 2019; Hutchinson et al., 2020). However, our understanding on medical corruption is still limited due to information unavailability. Many corrupt behaviours in the health sector are frequently hidden and reliable materials to investigate that corruption is scarce (Bergman et al., 2021; Weaver, 2021).

China, the largest transitional and developing country in the world, is also plagued by medical corruption. As China embarked on its market-oriented economic reform in the 1980s, healthcare facilities received few government subsidies and were left to survive on their own (Fu et al., 2017). The rigid price schedule that undervalued doctor services forced healthcare providers to seek profits from expensive drugs and consumables and opened up opportunities for medical corruption. Besides people’s experiences of giving red envelopes to doctors for better treatment, medical corruption frequently happened at the organizational level. As a result, public hospitals and pharmaceutical companies were often involved in scandals of briberies, embezzlement and insurance fraud (Zhang et al., 2014). In 2010, all public hospitals in Zhuhai city were involved in the crimes of receiving kickbacks; in 2013, the British pharmaceutical giant GlaxoSmithKline (GSK)’s bribery scandal shocked China (Martina, 2013)¹; in 2022, some employees of AstraZeneca China were found to tamper with genetic testing results for cancer patients in order to defraud medical insurance funds (Zheng, 2022). In addition, anecdotal evidence suggested that corruption exists widely in other areas of healthcare system, such as equipment procurement, infrastructure construction, performance assessment, personnel appointment and administrative approval of market entry. For example, the rapid expansion of infrastructures since 2009 had given hospital directors and officials opportunities to take bribes (Zhong, 2015). The lack of transparency and accountability mechanisms in Chinese health system further exacerbates these problems (Rose-Ackerman and Tan, 2014).

In recent years, the Chinese government made great efforts to combat medical corruption. Appendix Table S1 lists major interventions and reforms that may contribute to curbing medical corruption. The most notable one is the anti-corruption campaign that swept all sectors in China. It was initiated by President Xi Jinping after he assumed office in 2012, and it was enforced by the Central Commission for Discipline Inspection. It aimed to restore people’s trust and the legitimacy of the government by cracking down on corrupted civil servants and party members. The campaign had a large impact on the health sector. For example, Guangxi, one province alone, investigated over 4000 medical corruption cases between 2016 and 2020 (Zhang, 2021). Besides the anti-corruption campaign, the Chinese government implemented a series of health reform measures to reduce opportunities of medical corruption. For example, the national volume-based procurement (NVBP) was implemented to address the excessively high prices of drugs and consumables, which resulted in a range of price cut between 25% and 96% for 25 kinds of drugs (Yuan et al., 2021). In addition, the central government initiated reforms on the remuneration system for physicians and price adjustments for medical services. These reform measures aimed to raise physicians’ legitimate income and reduce their reliance on kickbacks.

Research on medical corruption in China is limited. Most of the existing studies focused on one area of medical corruption, and few studies have discussed the effectiveness of recent anti-corruption interventions (Rose-Ackerman and Tan, 2014; Shi et al., 2018; Li et al., 2022). Employing a mixed-methods design, this study aims to examine patterns of medical corruption in China, explore its key drivers and assess the perceived effectiveness of recent anti-corruption interventions. We study these issues by using extracted data of medical corruption cases from the China Judgement Online (CJO) website, by employing a nationally representative dataset from the China Health and Retirement Longitudinal Study and through interviews of 17 key informants. Our findings will not only inform healthcare researchers and policymakers to better understand medical corruption in China and develop future anti-corruption interventions but also provide lessons for other developing countries.

Methods

Conceptual framework

Drawing on works from Vian (2008), Gaitonde et al. (2016) and Onwujekwe et al. (2019), we develop our conceptual framework to analyse the medical corruption in China (as shown in Figure 1). We adapt the seminal framework from Vian (2008) to identify drivers of corruption among key actors in the health sector². The health sector is simplified into five key actors: regulators, payers, providers, suppliers and patients. Their corruption behaviours are mainly driven by three forces. The first force is pressure to abuse. Doctors and healthcare officials may get involved in medical corruption because their salaries are too low to make a living. Pressure may also come from colleagues and other actors involved in medical corruption. The second force is the opportunity to abuse power. For example, actors with monopoly and discretionary powers may have greater opportunities to abuse power for private gains. In contrast, an effective disciplinary system with a strict accountability and enforcement mechanism may reduce opportunities for medical corruption. The last force is rationalization. It refers to behavioural factors such as individual beliefs, attitudes and social norms. For example, some doctors may participate in medical corruption because they feel justified and that social norms support their decisions.

Figure 1.

Analytic framework

Notes: The framework is adapted from Vian (2008), Gaitonde et al. (2016) and Onwujekwe et al. (2019).

Following Gaitonde et al. (2016) and Onwujekwe et al. (2019), Figure 1 categorizes medical corruption into three major types: theft, bribery and misinformation. Theft is defined as taking resources without permission or right. Payers may embezzle money from public insurance funds. Bribery is defined as taking valued goods to influence a decision for private gain. Suppliers of pharmaceuticals, medical equipment and consumables may give kickbacks to healthcare providers to promote their sales. Misinformation is defined as falsifying information for private gain. Patients may collude with healthcare providers to defraud health insurance agencies. Appendix Table S2, which is adapted from Gaitonde et al. (2016) and Onwujekwe et al. (2019), documents the interaction between the actors and the corresponding types of medical corruption in China. These corrupt practices may affect health system performance through dimensions of efficiency, access, quality and public satisfaction.

Mixed-methods design

We used a mixed-methods design to investigate medical corruption in China, as shown in Appendix Figure S1. We first conducted the quantitative analysis that consists of two parts. We used extracted data from cases on CJO to examine the patterns of medical corruption. We also used household survey data from the China Health and Retirement Longitudinal Study (CHARLS) to examine the prevalence of informal payments. The results in quantitative parts provided useful information on the pattern and distribution of medical corruption for further qualitative analysis. We then conducted in-depth interviews with 17 key informants to corroborate our findings from quantitative analysis and explore the drivers of medical corruption in China, its impacts on health system performance and the perceived effectiveness of recent anti-corruption interventions.

Quantitative component

We analysed the patterns of medical corruption using extracted data of medical corruption cases on CJO website. The CJO is a government website that stores judgements from courts at all levels. Uploading judgements on this website is mandatory for all cases except for cases concerning national security, privacy infringement or juvenile delinquency³. The website was established in 2013 to promote judicial justice and credibility (Supreme People’s Court of China, 2013). By the end of 2021, more than 6.7 million judgement documents of criminal cases are available on the CJO website. These documents were widely used in the analysis of social, economic and criminal issues in China.

We constructed a dataset of medical corruption cases from the CJO to analyse characteristics and dynamics of crimes and offenders in medical corruption. Figure 2 shows the flow chart of identifying medical corruption cases from the CJO. We limited our search to the period between 1 January 2013 and 31 December 2019 and used a combination of criminal charges relevant to medical corruption and health system actors as our search terms.⁴ We obtained 10 003 cases from the search. Two researchers independently read these documents and identified cases of medical corruption. Only cases in which the defendant was from the health sector were included.⁵ To avoid double-counting, we excluded cases where only bribe-givers were defendants. Furthermore, we only kept the second instance of a case if both first and second instances were found. In the end, we obtained 3546 cases for our analysis.

Figure 2.

Flow chart of judgement document extraction

To extract data from the cases, two researchers read the judgement documents and independently coded the data. The corresponding author provided the tie-breaking opinion when there were controversies in the coding process. Criminal offender was used as the unit of observation. For each criminal offender, the following variables were extracted: age, gender, job title, occupation, types of corruption, areas in the health sector (drugs, consumables, equipment, infrastructure construction, administrative approval, etc.), the amount of money involved, the date of the first corrupt practice, trial date and the sentence. Depending on the occupation and job title of offenders, we further classified offenders by the type of actors (regulator, payer, provider and patient). Finally, we had a sample with 3892 offenders and 4140 crimes due to multiple offenders in one case and multiple crimes committed by one offender.

It should be noted that the CJO dataset only included three types of corruption crimes: embezzlement, bribery and insurance fraud. These three corruption crimes fell into corruption of theft, bribery and misinformation in the international literature, respectively. Other types of medical corruption in Appendix Table S2 were not covered in the CJO because they were not criminal behaviours in China’s judicial practice. For example, diversion of patients, informal payment and absenteeism were not commonly regarded as criminal, and thus, they were not covered in the CJO. Among them, informal payment is an important type of medical corruption that was frequently investigated in the previous literature (Schaaf and Topp, 2019; Liu et al., 2021; Dallera et al., 2022; Li et al., 2022). We used CHARLS data to investigate the prevalence of informal payment by inpatient and outpatient settings between 2011 and 2018. CHARLS is a nationally representative longitudinal survey for people over 45 years old and their spouses in China (Zhao et al., 2014).⁶ The CHARLS asked respondents whether they gave informal payment to healthcare providers when using inpatient and outpatient care.⁷ The data analyses were carried out using Stata 16.0, and the figures were plotted by Excel 2016.

Qualitative component

After finishing the quantitative analysis, we conducted semistructured phone interviews with key informants between October and November 2021. Following the heterogeneous purposive sampling, we selected participants from government agencies, public hospitals, social health insurance agency, pharmaceutical companies, patients and academia. Each participant was actively involved in China’s healthcare system. We generated a shortlist of potential respondents and contacted them for their consent to participate in the study. We recruited new participants until information saturation was achieved (Fusch and Ness, 2015)⁸. In the end, we sent out 21 invitations and 17 key informants agreed to participate. The characteristics of the interviewees are shown in Table 1. Among them, six participants were from public hospitals, three were drug sale representatives, two were government officials, one was a manager from social insurance agency, two were patients and three were scholars. Over 70% of participants were male, and nearly half of them were from Eastern China.

Table 1.

Characteristics of interviewed key informants

Characteristics	Description	N	%
Identity	Providers	6	35.3
	Suppliers	3	17.6
	Regulators	2	11.8
	Payers	1	5.9
	Patients	2	11.8
	Scholars	3	17.6
Gender	Male	12	70.6
	Female	5	29.4
Age groups	(20, 30)	2	11.8
	(30, 40)	4	23.5
	(40, 50)	6	35.3
	(50, 60)	4	23.5
	(60, +)	1	5.9
Education level	College and above	15	88.2
	High school	2	11.8
Regions	Eastern China	8	47.1
	Central China	7	41.2
	Western China	2	11.8

All interviews were conducted in Mandarin Chinese. An interview guide was drafted in Chinese and was translated into English (see Appendix Table S3). We asked respondents about the perceived prevalence of medical corruption in China, its drivers, its potential impacts on health system performance and the perceived effectiveness of recent anti-corruption interventions and healthcare reform measures. Data from these interviews were transcribed verbatim in Chinese. Two researchers read all transcripts twice and independently coded the dataset. The initial codes were deduced from the literature on medical corruption and then were refined after coding the first four transcripts (Vian, 2008; Shi et al., 2018; Onwujekwe et al., 2019). Then, the refined codes were applied to the rest of transcripts. We performed an inductive, thematic analysis to identify key categories related to prevalence, drivers, impacts and interventions of medical corruption (Braun and Clarke, 2006). Coding and analysis were carried out using the MAXQDA 2018 software.

Results

Types of medical corruption

The CJO dataset contained 3892 offenders involved in crimes of medical corruption. 77.5% of them were healthcare providers, 11.5% were regulators, 9.5% were payers and 1.5% were patients who defrauded health insurance agencies. These offenders were convicted of 4140 crimes in total. Appendix Figure S2 shows the sample distribution by the time of being arrested for criminals.

Table 2 shows the summary statistics of all the crimes. Bribery was the major form of medical corruption in China. It accounted for 68.1% of total corruptions. 79.8% of individuals who took bribes were from healthcare facilities, followed by governments (14.0%) and health insurance agencies (6.2%). The average monetary value of a bribe was 783 000 CNY (∼110 000 USD). However, the distribution of amount was skewed. The median of amount was only 207 000 CNY. In most cases, the amount of kickbacks physicians received was less than 100 000 CNY. However, 468 offenders who were officials or hospital directors received more than 10 million CNY. The largest bribe reached over 100 million CNY, and the offender was the director of a provincial hospital. In addition, most reported bribery crimes were undetected for more than 5 years before they were arrested. On average, offenders receiving bribes were sentenced to about 3 years of imprisonment. The punishment on physicians was less severe than officials. In some cases, physicians were relieved of criminal penalties if they confessed, which partly showed the leniency of the government to protect the scarce health workforce in China.

Table 2.

Summary statistics on the characteristics of corrupt behaviours

Mean (SD)	N	Percent	Amount received (¥CNY 10 K)	Age (years)	Female	Span of crime detection (years)	Sentence (years)
Briberies	2820	68.1	78.3 (246.9)	48.8 (7.0)	12.9% (33.5%)	5.9 (3.2)	3.2 (4.2)
Regulator	396	9.6	116.5 (314.7)	50.4 (6.2)	13.4%(34.1%)	6.6 (3.7)	5.2 (6.7)
Payer	174	4.2	95.5(542.8)	46.9 (7.6)	11.8%(32.3%)	6.2 (3.1)	3.4 (3.6)
Provider	2250	54.3	70.9 (191.0)	48.7 (7.1)	12.9% (33.5%)	5.7 (3.1)	2.9 (3.7)
Embezzlement	943	22.8	81.7 (270.9)	44.9 (8.8)	34.6% (47.6%)	4.9 (3.2)	2.9 (5.6)
Regulator	92	2.2	87.6 (264.9)	48.4 (7.2)	10.3% (30.6%)	5.5 (3.1)	4.8 (4.5)
Payer	121	2.9	153.4 (590.8)	42.6 (8.9)	28.8% (45.5%)	6.1 (3.5)	3.8 (9.6)
Provider	730	17.6	69.2 (166.5)	45.0 (8.8)	38.3% (48.6%)	4.6 (3.2)	2.6 (4.7)
Fraud	376	9.1	54.3 (145.0)	43.1 (9.6)	32.8% (47.3%)	4.6 (3.4)	2.8 (5.8)
Payer	120	2.9	65.8 (177.1)	39.5 (8.4)	31.1% (46.5%)	5.6 (3.6)	3.8 (9.4)
Provider	196	4.7	36.0 (94.1)	45.0 (10.1)	34.8% (47.8%)	3.5 (3.0)	2.2 (3.0)
Patient	60	1.4	90.1 (191.4)	44.0 (7.0)	35.0% (48.3%)	5.3 (2.8)	2.7 (2.3)

Notes: this table shows the characteristics of corrupt behaviours in our sample. N refers to the number of observations. If there are more than one type of corrupt behaviours in a conviction, we would count them in our analysis separately. Thus, the total number of crimes is 4140, larger than the number of offenders (3892) in our sample. Amount received refers to the total monetary values of the crime. In our sample, there are 248 defendants who have multiple crimes. Of these 248 defendants, a specific amount of each crime for 149 defendants can be distinguished. However, a specific amount of each crime for the remaining 99 defendants cannot be distinguished due to lack of information. For the 99 defendants, we use total amounts when calculating amounts for a specific crime. Span of crime detection refers to the years between the date when the defendant started corrupt practices and the date when the defendant was under arrest. Sentence refers to the years the defendant must stay in jail. The table reports sample means with standard deviations in parentheses.

To better understand the pattern of bribery crimes in China, we further showed the distribution of briberies by givers for each type of bribe-taker. As shown in Figure 3, 95% of total bribes that healthcare providers received were from suppliers, and the rest 5% were from providers. As for the bribes that regulators received, 67% were from suppliers and 29% were from providers. The staff of social health insurance agencies received bribes from providers and suppliers, accounting for 44% and 53%, respectively. A slight difference should be noted. In most cases, providers who bribed their superiors (public hospital directors or regulators) were staff of public hospitals for personnel promotion. Providers who bribed the payers were mainly private hospitals or primary health care institutions for the eligibility of social health insurance.

Figure 3.

Distribution of bribe-givers for each type of bribe-taker

Notes: Each subfigure shows the distribution of bribe-givers for each type of bribe-taker in our CJO sample. For example, Panel A shows that 95% of bribes the providers take were from suppliers, and only 5% were from providers. If there are more than one type of bribe-givers in a conviction, we would count them in our analysis separately. Thus, the total number of bribe-givers is larger than the total number of bribery crimes (2820) in Table 2.

Figure 4 further shows the distribution of bribery areas. It suggests that drug sales and equipment procurement are the primary areas of bribery, accounting for 32% and 26% of total bribes. This result was mainly driven by the fact that most individuals who took bribes were from healthcare facilities. Figure 4a shows that 36% and 29% of bribes the providers took were in the area of drug sales and equipment procurement and only 5% were for personnel recruitment and promotion. For regulators and payers who took bribes, their primary areas of bribery were different. As shown in Figure 4b and c, 22% of total bribes that regulators took were related to recruitment and personnel promotion and 76% of total bribes that payers took were related to administrative approval in the area of health insurance.

Figure 4.

Distribution of bribery purpose

Notes: Each subfigure shows the distribution of bribery areas for each type of bribe-taker in our CJO sample. Bribery can fall into six areas: drugs, equipment, consumable, infrastructure, insurance and personnel. Drugs means bribes given by drug companies or other relevant individuals. Equipment means bribes given by suppliers to promote the sales of medical equipment. Consumables means bribes given by consumable companies or other relevant individuals. Infrastructure means bribes given by suppliers of infrastructure construction. Insurance means that bribes are related to administrative approvals or other administrative activities in the area of health insurance. Personnel means bribery related to recruitment and personnel promotion. For example, Panel A shows that 36% of bribes the providers took were in the area of drug procurement or sale promotion, and only 5% were for personnel recruitment and promotion. If there were more than one area for a bribe-taker in a conviction, we would count them separately in our analysis. Thus, the total number of areas is larger than the total number of bribery crimes (2820) in Table 2.

These findings were confirmed by many interviewers. They mentioned that bribery was the main type of medical corruption in China. One interviewee’s words were echoed by other interviewees.

Commercial bribery is the main type of medical corruption in China. It often takes place in medical equipment procurement and sales of consumable materials and drugs. For example, pharmaceutical companies and sale representatives would bribe the deans of hospital departments, pharmacy directors and doctors to promote sales. My estimate is that the waste of medical resources due to medical corruption may account for more than 30% of the total healthcare expenditures in China. (P17, official, male, Eastern region)

Table 2 also reports the characteristics of embezzlement and fraud crimes. Embezzlement was the second most common type of corruption, accounting for 22.8% of total crimes. The average amount of an embezzlement crime was 817 000 CNY. 77.4% of individuals who embezzled funds were from healthcare facilities, 9.8% from governments and 12.8% from health insurance agencies. Its average span of crime detection was 4.9 years, slightly less than that of bribery. On average, offenders committing embezzlement were sentenced to 2.9 years of imprisonment.

Fraud crimes account for 9.1% of total crimes and were mostly insurance fraud. The average amount of fraud was 543 000 CNY. 52.1% of individuals who committed fraud were from healthcare facilities, 31.9% from health insurance agencies and 16.0% from patients. In some cases, primary health care institutions/private hospitals colluded with patients to produce false claims and commit medical insurance fraud, making the fraud harder to detect.

Table 3 reports the prevalence of informal payment among outpatients and inpatients aged 45 years old and above in recent years. It showed that between 2011 and 2018, the share of outpatients who gave informal payments to healthcare providers remained less than 1%. Inpatients were more likely to give informal payments than outpatients. The proportion of inpatients who gave informal payments exceeded 2% in 2011, and this number declined to 1.4% in 2018.

Table 3.

The prevalence of informal payment in China

	Outpatient		Inpatient
Year	n	% (95% CI)	n	% (95% CI)
2011	3354	0.7 (0.5-1.1)	1575	2.2 (1.6-3.1)
2013	3987	0.5 (0.3-0.7)	2381	2.3 (1.7-3.0)
2015	4079	0.3 (0.2-0.6)	2767	1.2 (0.9-1.7)
2018	3255	0.4 (0.2-0.7)	3323	1.4 (1.1-1.9)

Notes: The table shows the prevalence of informal payment in both outpatient and inpatient settings. CI refers to confidence interval. The data are from CHARLS, a nationally representative survey for people aged over 45 years old across 28 Chinese provinces.

Drivers of medical corruption

The respondents gave various reasons for the widespread medical corruption in China. Financial pressure and peer pressure were two reasons that were frequently mentioned by interviewed hospital directors and doctors. In addition, some officials and scholars emphasized institutional factors such as administrative oversight and anti-corruption enforcement. No respondents attempted to justify any corrupt behaviours in the interview.

Pressure to abuse

Providers face both economic and social pressures to abuse their power. Most doctors complained that salaries and bonus from hospitals were not enough so that they had to seek for other earnings. One doctor further mentioned that widespread prevalence of receiving kickbacks from drug and consumable might be attributed to the distorted price schedule of healthcare services and limited fiscal subsidies from governments. Two health workers said as follows:

In our county hospital, more than 60% of doctors cannot support their families if they solely rely on salaries and bonus from hospitals. (P14, hospital director, male, Central region)

The prices of health care services relying on labor inputs are relatively low. For example, the price of removing foreign objects from a patient’ eye is only 10 yuan, and the outpatient diagnosis fee is only 14 yuan per patient. But drugs and tests are lucrative. (P12, attending physician, female, Western region)

Some doctors and pharmaceutical sale representatives also believed that social pressure was another cause of widespread bribery among doctors. While junior doctors were often reluctant to be involved in bribery at first, they had risks to be isolated if they did not follow the corrupt practice of other doctors in their departments. One junior doctor noted that

When most doctors in my department receive kickbacks, I cannot refuse these kickbacks. I will be isolated if I don’t follow them. (P4, Attending physician, female, Central region)

When talking about the role of pressure in driving embezzlement, several hospital directors and officials denied this proposition. However, for some vulnerable hospitals, pressure to survive had pushed them into committing insurance fraud.

As far as I know, nobody was forced to embezzle money. High-ranking officials and directors have decent income, and they would not engage in embezzlement due to financial pressure. (P8, official from healthcare administration, male, Eastern region)

Because some small hospitals and private hospitals have trouble in making ends meet, they are prone to participate in medical insurance fraud and to make fraudulent insurance claims, forming a complete industrial chain of insurance fraud. (P6, official from local health security bureau, male, Central region)

Opportunity to abuse

Some respondent mentioned that embezzlement, bribery and insurance fraud in China’s health system were also driven by the dysfunctional internal control system and weak enforcement of anti-corruption interventions. It especially held true before the Xi administration came to power in late 2012. Even though a large-scale systematic anti-corruption campaign is running, enforcement needs to be improved, especially at the local level. The local Commissions for Discipline Inspection are sometimes constrained by manpower capabilities to conduct investigations, leaving rooms for corrupt practices in the health sector. One interviewed official’s words were echoed by other interviewees.

The internal supervision within an organization is still relatively weak in China, though it is stronger than before. If the top leader of an organization wants to be corrupt, he has numerous ways to make people in this organization to cooperate with him. The external supervision mechanism is also weak. The Commission for Discipline Inspection only starts an investigation if they receive clear clues. Some corrupt officials are not prosecuted due to a lack of investigation capacity. (P6, official from local health security bureau, male, Central region)

One scholar further commented that an important driver was bureaucratic interest. The healthcare administration, hospital directors and physicians had aligned interests in covering up corruption. On one hand, the healthcare administration was unwilling to hold public hospitals and their physicians accountable for medical corruption because public hospitals were often political supporters of the healthcare administration. On the other hand, local governments’ concern on economic developments also weakened their political will to combat bribe-giving behaviours by pharmaceutical companies and other business organization.

The health administration was unwilling to eliminate grey income such as kickbacks from pharmaceutical companies to protect the interests of hospitals and doctors. Local governments were reluctant to crack down on the corrupt behaviours of pharmaceutical companies, mainly because they were worried about its negative impacts on economic growth. Therefore, the lack of political will to combat medical corruption is one of the root causes. (P13, scholar, male, Eastern region)

In addition, irrationally high prices of drugs and consumables were also drivers for corruption, especially for bribe-taking behaviours. High profit margins left suppliers room for giving bribery and kickbacks. In some cases, pharmaceutical companies were competing for the amount of kickbacks to doctors. The amount of kickbacks had a significant influence on physicians’ treatment and prescription decisions. One pharmaceutical sale representative’s word is as follows:

In China, profit margins for drugs were extremely high and most doctors knew that. In some cases, we were competing for the amount of kickbacks to doctors. If we could give more kickbacks than other competitors, some doctors were willing to prescribe more our products. (P1, sale representative from a pharmaceutical company, male, Eastern region)

Rationalization to abuse

Almost all healthcare officials, hospital directors and doctors expressed the opinion that most corruption practices were forced to some extent. No respondents denied the illegality of medical corruption and its adverse impacts on health system performance. These words suggested that corrupt behaviours in the health sector were disgraceful in people’s mind, and social norms in China did not support such corrupt behaviours. The main drivers of medical corruption are pressure and opportunity rather than rationalization. A hospital director’s words were echoed by other interviewees.

Some corrupt behaviours were forced. Healthcare workers knew that their corrupt behaviours were shady and disgraceful, and it was a sensitive topic among them. But some of them had no choices because they had to survive. (P2, hospital director, male, Western region)

Impacts on healthcare systems

From various perspectives, respondents described negative impacts of medical corruption on health system performance. When talking about the corrupt behaviours of doctors, most respondents noted that it would increase health expenditure and reduce the accessibility of health services. One interviewed patient’s words are as follows, which were echoed by other interviewees,

The impacts of medical corruption on healthcare quality are unclear. However, kickbacks will incentivize doctors to over-prescribe and over-treatment, making healthcare services more expensive and inaccessible. (P5, patient, female, Central region)

One pharmaceutical sale representative emphasized that corruption practices would generate an adverse effect on China’s pharmaceutical industry in the long run. The business model of promoting sales by giving kickbacks may result in under-investments in research and innovation. In addition, one interviewed doctor expressed his concern that medical corruption intensified the tense doctor–patient relationship and indirectly resulted in violence against health workers. All these undesirable consequences could reduce the long-term supply of doctors. His words are as follows:

Medical corruption leads to a tense doctor–patient relationship. Many young students are unwilling to go to medical schools or decide not to be doctors after graduating from medical schools. (P3, Attending physician, male, Central region)

When talking about the impacts of corrupt practices by healthcare officials and hospital directors, most respondents mentioned that their corrupt behaviours would generate more harmful consequences than doctors. Specifically, an official from local health security bureau had witnessed that many doctors followed their corrupt director:

Corruption by healthcare officials and hospital directors would generate more harm. For example, a hospital director in our city received a large number of bribes and also made corruption behaviours widespread within the whole hospital. Many doctors and managers colluded with each other and engaged in corruption together. (P6, official from local health security bureau, male, Central region)

Perceived effectiveness of anti-corruption interventions

Most respondents focused on two anti-corruption interventions in the interview and expressed that they may be effective in reducing the prevalence of corrupt behaviours in the health sector. The first one was the anti-corruption campaign launched by the Xi administration since 2012. Especially, for officials and hospital directors, their opportunities to abuse power for private gains were largely suppressed. Their behaviours were frequently under scrutiny by anti-corruption inspection teams:

After the 18^th National Congress of the Chinese Communist Party, more supervision and restrictions were put on hospital directors and officials. There are many institutional changes. For example, information on bidding and procurement are required to be transparent and available online. The Commission for Discipline Inspection inspects public hospital more frequently. Undoubtedly, the anti-corruption campaign has obviously suppressed the rooms for medical corruption. (P2, hospital director, male, Western region)

The other measure that was frequently mentioned by respondents was the NVBP for drugs and medical consumables. It has substantially reduced prices of targeted drugs and medical consumables, leaving smaller room for kickbacks to doctors. Negotiations between healthcare security administration and pharmaceutical companies for new drugs had similar effects. However, doctors and sale representatives were candid that these measures did not eliminate medical corruption and corrupt behaviours still widely exist in more hidden ways. For example, one sale representative mentioned that kickbacks were given to some doctors for promoting the prescription of non-NVBP-list drugs and bribe-giving behaviours might be under the cover of other activities such as academic lectures and online healthcare services:

The centralized NVBP sharply cut down the prices of generic drugs, making kickbacks from pharmaceutical companies decrease substantially. (P8, official from healthcare administration, male, Eastern region)

The kickbacks still exist. Hospitals and doctors are incentivized to prescribe drugs that are not included in the NVBP list. Fewer sale representatives give kickbacks to doctors in cash. Instead, they are more likely to give kickbacks under the cover of academic promotion, such as lecture fees for experts and conference sponsorships. (P9, sale representative from a pharmaceutical company, female, Eastern region)

Three interviewees mentioned the special campaign against medical insurance fraud. The campaign mainly included unannounced inspections on healthcare providers, improvement in fraud detection with big data technology, collaboration with other administrative departments, etc. Overall, they thought that this campaign had promising effects on reducing medical insurance fraud, especially by private hospitals:

The special campaign can deter hospitals from engaging in medical insurance fraud. I participated in the campaign and conducted an investigation on several hospitals in my county. The amount of fraud in one private hospital was up to 2 million CNY. (P6, official from local health security bureau, male, Central region)

Some other measures to date were thought to have limited effects on reducing medical corruption. For example, three respondents (two doctors and one official) expressed the opinion that reforms on the remuneration system for physicians progressed slowly and physicians were still under financial pressure to receive bribes. For another instance, although payment reforms such as the implementation of Diagnosis-related Groups altered physician incentives to some extent, one respondent mentioned that doctors might take strategic behaviours to circumvent these reforms and their corrupt practices might remain largely unchanged.

Discussion and conclusion

By employing a mixed-methods design, this study investigated medical corruption in China. Using extracted data on court documents from the CJO website, we found that various forms of medical corruption existed in China; among them, bribery was the major form of medical corruption, accounting for 68.1% of total corruption cases in the health sector. More than 79.8% of bribe-takers were healthcare providers and most bribe-givers were suppliers of pharmaceuticals, medical equipment and consumables. In most cases, suppliers gave healthcare providers kickbacks to promote sales of their products. The patterns of medical corruption among governmental officials and payers were more complex. Briber-givers were more diverse, and motivation for bribery included personnel promotion, administrative approval and performance assessment. It is important to note that the patterns from the CJO data were under the full influence of the anti-corruption campaign, but only captured the initial impact of the NVBP and other anti-corruption measures. In other words, the patterns of medical corruption in China may have significant changes after 2020 because measures like the NVBP or the special campaign against medical insurance fraud would show their full effects then. Furthermore, the results based on the CHARLS suggested that the prevalence rate of informal payments from patients remained at a low level. Only 0.4% of outpatients and 1.4% of inpatients once ever gave ‘red envelope’ to physicians in 2018. Our finding is consistent with other existing studies in China (Liu et al., 2021; Li et al., 2022). For instance, Li et al. (2022) found that China’s informal payment rate in the inpatient setting in 2008 was 4.9%, which is slightly higher than the rate of 2.2 in 2011 in our study. However, the prevalence of informal payment in China is much lower than that in other low- and middle-income countries, which varies from 20% to 70% (Schaaf and Topp, 2019).

We conducted interviews with 17 key informants to explore the extent, effects and drivers of medical corruption in China. Consistent with the findings from quantitative analysis, they described that bribery was the main type of medical corruption in China, and it often took place in the procurement of medical equipment and sales of consumable materials and drugs. They emphasized that these corrupt behaviours would boost health expenditure growth, reduce the accessibility and affordability of healthcare services, and intensify the tense doctor–patient relationship. In the long run, medical corruption would also negatively influence the doctor supply and innovation in the health sector.

Interviewed hospital directors and doctors expressed that financial pressure was the main reason for bribe-taking behaviours among healthcare providers. This opinion pointed to the long-lasting perverse incentives faced by public healthcare facilities (Yip et al., 2010). They emphasized that their legitimate income could not reflect their labour input and make them lead decent lives, due to distorted prices of healthcare services and limited fiscal subsidies. To some extent, they were ‘forced’ to receive bribes and kickbacks from medical suppliers. In addition, some governmental officials and scholars attributed medical corruption in China to the weak oversight and weak enforcement of anti-corruption. This driver was especially prominent before Xi Jinping assumed office in late 2012. They argued that even after 2012, capacity constraints and bureaucratic interests were two potential impediments to combat medical corruption in China. These drivers were consistent with findings in earlier studies of anti-corruption and healthcare reform in China (Hsiao, 2007; Zhu, 2015; Li and Wang, 2019).

When talking about the effectiveness of recent anti-corruption interventions, most respondents pointed to the anti-corruption campaign launched by the Xi administration since 2012. They expressed that this campaign not only investigated and punished many corrupt officials and doctors but also suppressed opportunities to abuse power through governance strengthening and institutional changes. Under stricter oversight, healthcare officials and hospital doctors would not dare to receive bribes as before. Their words suggest China’s positive progress in combating medical corruption, which is echoed in other sectors as well (Deng, 2018). In addition, many interviewees mentioned the effectiveness of the NVBP in reducing kickbacks from pharmaceutical companies. The special campaign against medical insurance fraud was also perceived to have promising effects on suppressing medical insurance fraud. However, other healthcare reform measures, such as payment reform, price changes of medical services and public hospital performance evaluation, were perceived to have limited effects on corrupt behaviours. Some existing evidence supports the aforementioned findings on the perceived effectiveness of recent anti-corruption interventions in China. For example, the Corruption Perceptions Index of China significantly increased from 36 to 42 between 2014 and 2020, suggesting the general effectiveness of the anti-corruption campaign launched by the Xi administration. In addition, among listed pharmaceutical companies with annual revenues of more than 10 billion CNY in 2020, the share of sales expenses in total expenses decreased from 18% to 15% between 2019 and 2021, suggesting the effectiveness of recent anti-corruption interventions such as the NVBP in suppressing kickbacks.

To the best of our knowledge, this is the first study that investigates the patterns of medical corruption in China and examines its key drivers and impacts on health system performance. The findings in this study can help policymakers and researchers reach consensus on ways to develop evidence-based anti-corruption interventions. Moreover, this study is a response to recent international calls for deepening relevant research on medical corruption (García, 2019; Hutchinson et al., 2019; Vian, 2020). China’s case allows global researchers to have a richer understanding of medical corruption in developing countries. For instance, the widespread provider behaviours of taking bribes from suppliers in China suggest the necessity of regulating physician–industry interactions, which is an important lesson from Europe and the USA (Gaitonde et al., 2016; Vian, 2020). Previous studies on medical corruption in developing countries focus more on bribery from patients or informal payments but pay less attention to this type of corruption (Mackey et al., 2018; Onwujekwe et al., 2019; Schaaf and Topp, 2019).

Our findings of interventions’ perceived effectiveness contribute to the literature where a paucity of evidence exists regarding how best to reduce corruption (Gaitonde et al., 2016; Rispel et al., 2016; Onwujekwe et al., 2019). Referring back to Vian (2008)’s framework, the perceived effectiveness of anti-corruption campaign and NVBP in China highlights the importance of governance strengthening and incentive realignment in suppressing opportunities and pressures for medical corruption. China’s case also suggests the vital role of strong political will and leadership in combatting corruption and facilitating systemic reforms (Huss et al., 2011; Ankamah and Manzoor E Khoda, 2018; Gilson and Agyepong, 2018). However, this study provides little evidence on the effectiveness of other commonly recommended anti-corruption measures in the international literature, such as increased transparency, whistleblowing and complaint mechanisms, because they are not key components of anti-corruption measures in China.

It should be noted that corrupt behaviours in the Chinese health system still widely exist, as suggested in the interviews. Some new forms of bribe-giving behaviours emerge under the cover of other activities, such as academic conferences and provision of online healthcare services. More strategies are in need to achieve a sweeping victory. Our first recommendation is to enhance administrative oversight. As stated by interviewees, capacity constraint is an impediment to combat medical corruption in China. China may fill this gap through establishing stronger information and audit systems. For example, the tools of big data analysis can monitor the financial relationship between physicians and suppliers and can provide clues for investigation on kickbacks and bribes. These systems can also be used to improve transparency of healthcare officials’ decisions and doctors’ prescription behaviours. They can contribute to increasing public engagement in scrutiny of public actors and their decisions. Lessons from other sectors, such as the civil services, suggest that information and audit systems can play an important role in fighting corruption (Kim et al., 2009; Zhao and Xu, 2015). Public engagement is another necessary factor in enhancing administrative oversight, among which whistleblowing is playing an important role. Whistle-blowers may combat medical corruption by sending anonymous or real-name tips to Commission for Discipline Inspection or disseminating information on social media like Weibo. Although internal whistleblowing has been officially encouraged in China, citizens have a low willingness to act as whistle-blowers (Su, 2020). Establishing programmes to further protect and reward whistle-blowers are essential to encourage whistleblowing (Vian, 2020).

Second, a systemic reform of public hospitals should be taken. Although the anti-corruption campaign and the NVBP suppress the opportunity for medical corruption, financial pressure to take kickbacks still hold among many healthcare providers, especially public hospital doctors. Some Chinese doctors are ‘forced’ to take desperate measures under the financial pressure. Although the central government has pushed local governments to reform physicians’ remuneration system by substantially increasing legitimate income and strengthening performance evaluation since 2016, its progress is relatively slow. Local governments have concerns that such reforms may increase fiscal expenditures and incur huge social costs such as massive complaints from senior doctors and compromise in quality of care. The Sanming model in China, a successful example, suggests that the success of a remuneration system reform for physicians relies on the alignment of various reform measures, including reforms on hospital governance, internal organization, management, pricing of healthcare services and payment methods (Fu et al., 2017). In the future, the central government should issue mandatory guidelines and set explicit goals for local governments to implement such reforms. When social reforms entail distributive conflicts and bureaucratic friction, top-down pressure for compliance is a dominant driver for local adoption in China (Huang and Kim, 2020). The local governments should mobilize financial and manpower resources to facilitate the implementation.

Finally, to further curb kickbacks and bribes, we recommend mandatory information disclosure of payments from suppliers to physicians in China. In 2010, the Physician Payments Sunshine Act was enacted in the USA, and it forced drug and medical device firms to publicly disclose interactions with physicians (Agrawal et al., 2013; Chao and Larkin, 2022). Since then, many developed countries like Japan, France and Australia implemented similar ‘Sunshine’ policies ‘to improve transparency in the health sector’ (Grundy et al., 2018). It is considered as necessary, though not a sufficient condition for prevention of medical corruption (Lexchin and Fugh-Berman, 2021). In addition, China may consider increasing punishment for bribe-givers. In mainland China, bribe-takers are the main targets of anti-corruption agencies and their criminal justice system. Bribe-takers are punished more severely than bribe-givers. Theoretically, bribe-givers may be liable to criminal prosecution. But, in practice, they often receive a lenient sentence, and may even be exempted from criminal penalties if they confessed and cooperated. Liu and Wang (2015) concluded that this asymmetric punishment mechanism in China has induced more bribe-givers and led to the spread of commercial bribery.

Some limitations of this study should be noted. First, the CJO dataset used in this study is a subsample of medical corruption in China. Admittedly, many cases of medical corruption are not detected, and sample selection may be a concern for this study. However, this dataset is the best available one for this study. Second, the CHARLS dataset only covers people aged over 45 years old in China. Given that older people have more severe conditions than younger people in general, they are more likely to give ‘red envelope’, and thus, the prevalence of informal payment in this study may be overestimated. Even so, we find a low prevalence rate of informal payment. Third, we do not include all types of medical corruption in our analysis. For example, although they are often included in the international literature, absenteeism and patient diversion are usually not regarded as crimes of medical corruption by the Chinese society. Forth, due to the sensitivity of this topic, key informants were likely to conceal some of their thoughts and facts in the interviews. It may result in an underestimate in the severity of medical corruption in China, an overestimate in the effectiveness of anti-corruption interventions and omissions of some other drivers. These limitations should be addressed in future research.

Supplementary data

Supplementary data are available at Heath Policy and Planning online.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author. Please send data request to [email protected] if needed.

Funding

This study was financially supported by the National Social Science Fund of China (grant number 20ZDA075).

Acknowledgements

We would like to thank Ling Li, Ming Wu, Qiulin Chen, Duo Xu and seminar participants at Peking University for their helpful comments. The opinions expressed in this article and any errors are those of authors alone.

Author contributions

H.F. designed this study. Y.L. and Y.L. collected the data and conducted the analysis. H.F. and Y.L. wrote the first version of the manuscript. All authors interpreted the findings and edited the manuscript. All authors approved the final draft for publication.

Reflexivity statement

The authorship team intentionally reflects diversity and inclusion in gender (three females and two males), seniority (two PhD candidates, one post-doctoral scholar, one assistant professor and one senior professor) and region (three from East Asia and two from North America). Two authors have extensive experience conducting qualitative fieldwork in China, and all the authors have extensive experience in conducting quantitative analysis.

Ethical approval

Ethical approval is not required for this study.

Conflict of interest

None declared.

Endnotes

1.

In 2013, the GSK was accused of engaging in ‘massive and systemic bribery’ to hospitals, doctors and government officials to promote drug sales. The total amount of bribery over 6 years was estimated to be more than 3 billion Yuan. In 2014, after a 15-month investigation led by the Chinese government, Mark Reilly, GSK’s former top executive in China, pleaded guilty to bribery-related charges and was given a three-year prison sentence, suspended for 4 years. Four other senior Chinese GSK managers also received suspended sentences, and 312 employees in China were dismissed. As a consequence, GSK was fined 3 billion Yuan ($488 million), the largest corporate fine in China. More information on this scandal is available at https://en.wikipedia.org/wiki/GSK_China_scandal.

2.

Based on ideas of neoclassical economic theory and good governance, Vian (2008)’s framework conceptualized corruption as a result of individual actions and systems-level problems. This framework was widely applied to guide country-specific analysis of medical corruption (Huss et al., 2011; Rispel et al., 2016; Onwujekwe et al., 2019). For example, Rispel et al. (2016) applied this framework to explore the pervasive medical corruption in South Africa. Onwujekwe et al. (2019) adapted the framework to identify factors sustaining different forms of medical corruption in West African. Vian (2020) gave a comprehensive overview on the literature, which applied the framework.

3.

Court documents before 2014 were not mandatory for posting. Courts may post them at their discretion.

4.

Specifically, we used the search terms ‘bribery or embezzlement or corruption or kickbacks or fraud’ and names of health system actors (like ‘Hospitals’ or Health Insurance’) to extract relevant judgement documents on medical corruption.

5.

Some corrupt behaviours happened in healthcare facilities, but offenders were not from the health sector. For instance, some officials in other sectors took bribes when they or their relatives were in hospital. These cases were not medical corruptions and should be excluded in our sample.

6.

Although CHARLS does not cover people aged under 45 years, it can provide us the information on the prevalence of informal payments in China. This is because people aged above 45 years account for more than 70% of inpatient service utilization in China.

7.

For example, in the wave of CHARLS 2013, the respondents were asked ‘did you give any red envelopes to the doctors for this outpatient visit?’ (question ED029) and ‘did you give any red envelopes to the doctors for this hospital admission?’ (question EE032).

8.

We relied on the feasibility of new coding to see whether data saturation was reached. When further coding was no longer feasible, we concluded that our study had reached data saturation.