Prevalence and determinants of defensive medicine among physicians: a systematic review and meta-analysis

Abstract

Defensive medicine, characterized by physicians’ inclination toward excessive diagnostic tests and procedures, has emerged as a significant concern in modern healthcare due to its high prevalence and detrimental effects. Despite the growing concerns among healthcare providers, policymakers, and physicians, comprehensive synthesis of the literature on the prevalence and determinants of defensive medicine among physicians has yet been reported. A comprehensive literature search was conducted to identify eligible studies published between 1 January 2000 and 31 December 2022, utilizing six databases (i.e. Web of Science, PubMed, Embase, Scopus, PsycINFO, and Cochrane Library). A meta-analysis was conducted to determine the prevalence and determinants of defensive medicine. Of the 8892 identified articles, 64 eligible studies involving 35.9 thousand physicians across 23 countries were included. The overall pooled prevalence of defense medications was 75.8%. Physicians engaged in both assurance and avoidance behaviors, with the most prevalent subitems being increasing follow-up and avoidance of high-complication treatment protocols. The prevalence of defensive medicine was higher in the African region [88.1%; 95% confidence interval (CI): 80.4%–95.8%] and lower-middle-income countries (89.0%; 95% CI: 78.2%–99.8%). Among the medical specialties, anesthesiologists (92.2%; 95% CI: 89.2%–95.3%) exhibited the highest prevalence. Further, the pooled odds ratios (ORs) of the nine factors at the individual, relational, and organizational levels were calculated, and the influence of previous experience in medical-legal litigation (OR: 1.65; 95% CI: 1.13–2.18) should be considered. The results of this study indicate a high global prevalence of defensive medicine among physicians, underscoring the necessity of implementing targeted interventions to reduce its use, especially in certain regions and specialties. Policymakers should implement measures to improve physicians’ medical skills, enhance physician–patient communication, address physicians’ medical-legal litigation fears, and reform the medical liability system. Future research should focus on devising and assessing interventions to reduce the use of defensive medicine and to improve the quality of patient care.

Introduction

Defensive medicine has become a prominent issue in contemporary healthcare, marked by physicians’ tendency to order excessive diagnostic tests, procedures, or referrals and avoid certain patients and procedures [1, 2]. It is often motivated by concerns about litigation and malpractice claims rather than by patients’ best interests, potentially decreasing the quality while increasing the costs of patient care [3]. Previous studies indicate that defensive medicine can be divided into assurance and avoidance behavior [4, 5]. Assurance behavior is regarded as a positive defensive medicine that prevents unnecessary and additional medical treatments [6], whereas avoidance behavior is considered a negative defense behavior. Owing to the fear of becoming involved in medical litigation, physicians implement avoidance behaviors by avoiding high-risk procedures or patients [7]. These phenomena have been linked to various negative outcomes, including increased healthcare costs, reduced quality of care, and patient harm [8–10].

The prevalence of defensive medicine significantly varies across countries and specialties. A comprehensive national study revealed that 92% of physicians in the USA engaged in defensive medicine practices [10]. In contrast, European data indicated that 74.4% of Belgian physicians and 45.39% of Spanish physicians reported employing such practices [11, 12]. Moreover, physicians of different specialties were found to have different degree of defensive medicine utilization. Several factors have been identified as the determinants, including practice type [4], previous experiences with medical-legal litigation [13, 14], and pressure from patients or colleagues [13]. Moreover, healthcare system factors such as reimbursement policies, quality metrics, and regulatory requirements may also contribute to the prevalence of defensive medicine [3, 15].

To date, there has been no comprehensive synthesis of existing literature on the prevalence and determinants of defensive medicine among physicians. By identifying the prevalence and determinants of defensive medicine among physicians, this systematic review aimed to inform the development of targeted interventions to reduce defensive medicine practices and promote high-quality patient-centered care.

Methods

Data sources and search strategy

As defensive medicine is a multifaceted phenomenon involving medicine, law, and society, multiple electronic literature databases were used. In accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [16], we conducted a comprehensive literature search spanning from 1 January 2000 to 31 December 2022 across six major and comprehensive databases: Web of Science, PubMed, Embase, Scopus, PsycINFO, and the Cochrane Library. A broad search strategy was used to minimize the risk of bias. Based on the standardized Medical Subject Headings term ‘defensive medicine’ in the PubMed database, four groups of keywords were generated. The full search strategy is shown in Supplemental File 1.

EndNote software (version 20) was employed to manage the citations.

Eligibility criteria

The following eligibility criteria were defined: (i) defensive medicine was performed by or related to practicing physicians from any specialty; (ii) the concept of defensive medicine was clearly defined; (iii) the prevalence of defensive medicine among physicians was explicitly reported, including both overall and subitems; (iv) studies that reported defensive medicine among physicians not in emergencies or extreme circumstances (such as warfare or pandemic); (v) original studies being published in peer-reviewed journals; (vi) full text was available; and (vii) for duplicate publications, the most extensive studies were included.

Study selection and data extraction

Two investigators independently screened the titles and abstracts generated by the search strategy. Once one of the investigators determined that an abstract was eligible, a full-text article was obtained. Subsequently, two investigators examined the full texts and assessed the final study inclusion by referring to the eligibility criteria. Disagreements were resolved by a third investigator. To determine whether additional articles should be included, the reference lists of the included studies were manually searched. Finally, all investigators participated in data extraction.

Information about defensive medicine and its determinants was extracted from each included study using a standardized form (Table 1). Descriptive details included the first author, year of publication, research location, medical specialty investigated, sample size, outcome dimensions, influencing factors, and definition of defensive medicine. The authors of the included studies were contacted for additional clarification or data if needed.

Quality assessment

The quality of the included studies was assessed using the Joanna Briggs Institute Critical Appraisal Checklist for Studies Reporting Prevalence Data, updated in June 2022 [17, 18]. Two investigators independently assessed study quality, and any disagreements were resolved by consensus.

Data synthesis and analysis

The main outcomes of this review were the prevalence and determinants of defensive medicine among physicians. In the included studies, the pooled prevalence of overall defensive medicine and its subitems was calculated with a 95% confidence interval (CI). For studies that did not report the overall prevalence of defensive medicine, we used the subitem with the highest prevalence rate as a proxy for the total to obtain a conservative estimate. Considering the expected heterogeneity among the included studies owing to variations in geographical location, population, specialty, and culture, we utilized a random-effects model. Subgroup analyses were also performed to compare the pooled prevalence of defensive medicine between studies conducted in different regions of the world and among physicians from different specialties.

We further performed a meta-analysis of the determinants of defensive medicine among physicians, including only factors examined in at least two studies. In the meta-analyses, the pooled odds ratios (ORs) and 95% CIs were reported to assess the effect of the factors, and forest plots were used to visualize the variability among studies. Finally, funnel diagrams and Egger’s tests were used to assess the publication bias in the studies included in the meta-analysis. A significance level of 5% was used for this review. All analyses were performed using Stata 16.0 and the R project.

Results

Search results and included studies

Search results

A systematic literature search initially yielded 8892 records, from which 64 eligible studies were included in review. The PRISMA diagram of the systematic search is shown in Fig. 1, and references for the studies included in this review are presented in Supplemental File 2.

Characteristics of the studies included

Through a full-text review, we extracted contextual details from the included studies (Table 1). The results revealed that all 64 studies encompassing over 23 countries and scrutinizing more than 35 900 physicians reported the overall prevalence of defensive medicine and/or its subitems. Among these studies, 20 were conducted in the European region, which ranked first. The prevalence of defensive medicine across different specialties was also explored, with surgeons receiving the most attention (n = 24). All included studies utilized a cross-sectional design. Moreover, 37 studies discussed the determinants of defensive medicine among physicians. All included studies were rated as having acceptable quality according to the quality assessment, as shown in Supplemental File 3.

Definition and measurement of defensive medicine in the studies included

Based on a comprehensive review of the definitions of defensive medicine in the included studies (listed in Table 1), three fundamental elements of the concept were observed: (i) purpose—aiming to avoid being subject to lawsuits, criticisms, or disputes; (2) process—the alteration of medical practice; and (3) result—a deviation from patient-centered care and the best interest of the patient. Out of the 64 included studies, 16 studies employed a binary categorization to assess defensive medicine, where ‘no’ signified the absence of defensive medical behavior. Meanwhile, 33 studies employed multi-dimensional Likert scales, encompassing terms with negative connotations like ‘never’ and ‘rarely’ to indicate the absence of defensive medical behavior. Additionally, 15 studies evaluated defensive medicine by asking respondents to specify the frequency of such behavior, with only ‘none’ denoting the absence of defensive medical behavior.

Prevalence of defensive medicine among physicians

Prevalence of defensive medicine across regions

Our review examined the total prevalence of defensive medicine among physicians globally. The overall pooled prevalence of defensive medicine was 75.8% (95% CI: 72.3%, 79.3%), as shown in Supplemental File 4. Notably, the prevalence of defensive medicine varies considerably across countries, and the pooled prevalence in each country included in this review is listed in Supplemental File 5. To provide a comprehensive view of the data, we created a hierarchical map, illustrating the prevalence of defensive medicine in various countries (shown in Fig. 2).

To further explore the regional differences in the prevalence of defensive medicine, we calculated regional estimates by grouping countries according to the six regions defined by the World Health Organization. Simultaneously, we categorized countries into different income groups based on the World Bank classification to assess the variance of defensive medicine across countries of different economic levels. The pooled prevalence of defensive medicine was relatively high in Africa (88.1%; 95% CI: 80.4%–95.8%), the Americas (80.8%; 95% CI: 76.8%–84.9%), and the Western Pacific region (76.1%; 95% CI: 64.0%–88.3%). Comparatively, the prevalence was lower in European (70.8%; 95% CI: 64.3%–77.2%) and Eastern Mediterranean regions (73.2%; 95% CI: 62.6%–83.8%). Notably, the lower-middle-income group (89.0%; 95% CI: 78.2%–99.8%) exhibited a higher prevalence than the upper-middle-income group (77.3%; 95% CI: 68.6%–85.9%) and high-income group (74.6%; 95% CI: 70.8%–78.5%), which is shown in Fig. 3.

Prevalence of defensive medicine by specialty

According to the subgroup analysis by specialty, the prevalence of defensive medicine varied markedly among different medical professionals. Specifically, anesthesiologists (94.4%; 95% CI: 84.1%–104.6%), medical technicians (89.1%; 95% CI: 84.6%–93.7%), general practitioners (83.5%; 95% CI: 69.4%–97.7%), and surgeons (77.8%; 95% CI: 73.5%–82.0%) exhibited a higher prevalence of defensive medicine than the overall pooled prevalence. In contrast, the prevalence rates of defensive medicine among internists, obstetricians/gynecologists, psychiatrists, and pediatrics were lower than the overall pooled prevalence, as shown in Fig. 3. It is worth noting that emergency medical specialists exhibited a slightly higher prevalence than the overall pooled prevalence.

Prevalence of defensive medicine subitems

The frequency of practicing the subitems of defensive medicine varied, as shown in Fig. 4. The subitem with the highest prevalence rate in the category of assurance behavior was increasing follow-up for defensive purposes, which was practiced by 79.8% of the physicians. Moreover, there were two other assurance behaviors with a prevalence of over 60%: consulting specialists (without transferring the primary responsibility of the patient’s care) more often than medically indicated (68.5%; 95% CI: 52.3%–84.7%) and ordering more tests or images than medically indicated (62.2%; 95% CI: 56.0%–68.3%). In contrast, the least frequently practiced assurance behavior was performing invasive procedures contrary to professional judgment (35.6%; 95% CI: 26.5%–44.7%). Furthermore, physicians also made unnecessary referrals (50.8%; 95% CI: 41.0%–60.5%) and prescribed more medications than medically indicated (46.2%; 95% CI: 34.6%–57.8%).

The most adopted avoidance behavior was avoidance of treatment protocols, with high-complication rates (72.3%; 95% CI: 61.6%–82.9%). Physicians also engaged in avoiding or eliminating certain procedures (44.5%; 95% CI: 31.6%–57.3%) and avoiding patients with specific conditions (45.5%; 95% CI: 37.3%–53.7%).

Determinants of defensive medicine among physicians

A typology of the determinants

Among the included studies, 38 studies investigated the determinants of defensive medicine using quantitative or mixed-methods approach. Detailed information on these determinants is presented in Supplemental File 6. They can be classified into individual, relational, organizational, and environmental levels.

Results of the meta-analyses for determinants of defensive medicine

Nine determinants discussed in 21 articles were used in the meta-analysis. At the individual level, physicians with more than 20 years of working experience (OR: 0.82; 95% CI: 0.55–1.09), being 50 years old or above (OR: 0.68; 95% CI: 0.33–1.03), and with a senior technical title (OR: 0.68; 95% CI: 0.44–0.93) were associated with less likelihood of practicing defensive medicine (Fig. 5). In addition, physicians in group practice had a lower likelihood of defensive medicine than those in solo practice (OR: 0.81; 95% CI: 0.59–1.04). At the relational level, the pooled results indicated that physicians having previous experience of medical-legal litigation (OR: 1.65; 95% CI: 1.13–2.18) or with colleagues having previous experience of medical-legal litigation (OR: 1.31; 95% CI: 1.00–1.61) were more likely to conduct defensive medicine. At the organizational level, physicians working in tertiary hospitals tended to have fewer defensive medical practices (OR: 0.57; 95% CI: 0.40, 0.74).

Notably, no environment-related determinants met the inclusion criteria for the meta-analysis. Among the studies analyzed, environment-related determinants were either unmeasured or reported inconsistently, precluding us from including them in the examination.

Publication bias

The results of our systematic review, as shown in Supplemental File 7, suggested minimal publication bias in the included studies, except for a slight bias in the working experience variable (Egger’s test P-value:.027). Nonetheless, the overall results remained robust and reliable, owing to the minimal publication bias observed in most studies. An evaluation of publication bias took into consideration the quality of the included studies. As detailed in Supplemental File 3, each study encompassed in this review has been assessed and deemed to meet acceptable quality standards.

Discussion

Statement of principal findings

This systematic review and meta-analysis revealed an overall pooled prevalence of 75.8% for defensive medicine use among physicians, underscoring the global severity of the issue. This has substantial implications for patient care, especially for addressing increased healthcare costs and potential harm to patients [4]. Disparities in the prevalence of defensive medicine were observed across countries, regions, and specialties. Additionally, factors at the individual, relational, organizational, and environmental levels can have an impact on defensive medicine.

Interpretation within the context of the wider literature

Defensive medicine was more prevalent in Africa, the Americas, and the Western Pacific than in European and the Eastern Mediterranean regions. The observed differences could be a result of the disparities in the cultural contexts in which healthcare settings are based (e.g. blame culture) [19]. We also found that the defensive medicine was more prevalent in low- and middle-income countries compared to high-income countries, which may be attributed to three factors. First, physicians in high-income countries are often required to participate in continuing medical education programs to maintain their licenses, equipping them with the most up-to-date knowledge of the best medical practices [20, 21]. Second, alternative dispute resolution methods, including communication and resolution programs in the USA [22] and Canada [23], as well as disclosure and apology laws [24], promote physician–patient communication, and avoid litigation [25]. Third, robust medical liability legal systems have been implemented, such as no-fault compensation systems in New Zealand [26], Denmark [27], and Finland [28]; tort reform in the USA [29]; and stronger systems of medical liability insurance [11, 12], which provide physicians with a sense of safety toward potential medical liability risks.

Variations in the prevalence of defensive medicine were observed across medical specialties, with anesthesiologists and medical technicians exhibiting the highest adoption rates. Notably, general practitioners ranked third in terms of defensive medicine prevalence, and as a result, more effective support is needed to empower gatekeepers to prevent or reduce the overuse of medical resources in primary healthcare units [30]. Defensive medicine in surgery received the most attention in the studies included in our review, especially in neurosurgery. These studies emphasize the importance of implementing interventions to mitigate defensive medical practices among physicians who perform high-risk procedures daily [15]. Our findings suggest that policy interventions should be tailored to the specific needs of different physician specialties [31]. For instance, the implementation of evidence-based guidelines that furnish obstetrician-gynecologists with a well-defined framework for labor and delivery has been recognized as a useful measure to reduce defensive medicine [14, 32]. For anesthesiologists and medical technicians, it is important to provide regular education and training on new technologies and equipment. Additionally, educating and training physicians on informed consent and shared decision-making could help reduce the probability of malpractice claims, particularly in psychiatry [33, 34].

The meta-analysis indicated that physicians with less experience, working in solo practice, or holding junior titles are more likely to engage in defensive medicine. Therefore, more attention should be paid to designing targeted interventions and support systems for these specific groups [35], like enhancing their medical technology proficiency, which could reduce their ordering additional laboratory tests or imaging, or making more non-essential referrals as a strategy to mitigate potential legal liabilities or risks [36, 37]. It is also important to acknowledge that, as mentioned in a few studies included, there can be a contrasting trend as well [7, 14]. Specifically, some experienced physicians could also become more cautious, leading to an increase in defensive medicine, especially in areas with high litigation rates. This highlighted the complexity of the relationship between work experience and defensive medicine. The findings also revealed that physicians with a history of medical-legal litigation or those who had witnessed colleagues facing such litigation were more likely to practice defensive medicine. Policymakers should design measures to reduce the fear of malpractice litigation among physicians. This can be achieved by enhancing communication between physicians and patients through patient-centered communication techniques [38]. Efforts should also be directed toward increasing transparency in medical decision-making [39] and implementing evidence-based clinical practice guidelines to reduce variability in medical decision-making [39, 40]. Additionally, interventions should be aimed at cultivating a supportive professional environment and encouraging a culture of open communication and learning from errors rather than focusing on blame and punishment [41]. At the organizational level, physicians employed in tertiary hospitals exhibited a lower level of defensive medicine. Accordingly, defensive medicine practice by physicians in clinics, primary and secondary hospitals should receive more attention.

We conducted a qualitative investigation of the environment-related determinants. The high prevalence of defensive medicine within the institutional environment may be attributed to the malpractice crisis level or medical-legal risk environment, premium burden, and coverage of medical liability insurance [19, 42]. Additionally, physicians in countries with complex legal systems and higher malpractice insurance premiums were more likely to practice defensive medicine. In countries with a high prevalence of defensive medicine, the legal system often requires high burden of proof from physicians, making it difficult for them to defend against malpractice claims [43, 44]. In contrast, countries with a lower prevalence, such as Spain, featured a legal system favoring physicians [45]. One feasible policy intervention is to introduce a medical liability system and tort reform [46], including measures such as capping noneconomic damages, streamlining the investigation and litigation processes for medical malpractice claims, and introducing alternative dispute resolution mechanisms such as arbitration or mediation [47, 48]. Tort reform has demonstrated effective reduction of defensive medicine in some states in the USA [49]. Blame culture in healthcare settings also contributes to the prevalence of defensive medicine. Public education campaigns addressing healthcare complexities and promoting realistic patient expectations regarding treatment outcomes may help reduce dissatisfaction and the likelihood of malpractice claims [13, 19]. Moreover, fostering responsible and accurate media coverage of medical malpractice issues could assist in shifting societal attitudes toward a more balanced perspective on litigation and compensation [3, 25].

Implications for policy, practice, and research

Interventions should be tailored to the characteristics of different physician specialties and region, addressing the underlying factors that contribute to defensive medicine. Various measures can be implemented, including enhancing physicians’ medical technology proficiency, improving communication between physicians and patients, reducing the fear of litigation, and establishing stronger medical liability insurance systems. Efforts should also be directed toward addressing the social factors that contribute to blame culture and promoting defensive medical practices. Future research should focus on devising and evaluating the efficacy of these interventions to reduce the use of defensive medicine and to enhance the quality of patient care.

Strengths and limitations

This is the first review to analyze the global prevalence and determinants of defensive medicine, highlighting the heterogeneity of prevalence and the effectiveness of measures. This review also had some limitations. First, the included studies used various measurement tools to assess defensive medicine, which may have affected the comparability of the results. Second, the subitem with the highest prevalence rate was used, which inevitably led to an underestimation of the actual prevalence rate. Finally, all included studies used self-reported measures of defensive medicine, which may contain reporting bias.

Conclusion

The high prevalence of defensive medicine among physicians imposes a significant burden on healthcare systems, with potentially negative consequences for both patients and practitioners. Additionally, the necessity for policy interventions was emphasized, particularly for physicians in specialties with high-risk procedures and litigious environments.

Acknowledgements

Not applicable.

Supplementary data

Supplementary data is available at INTQHC online.

Conflict of interest

No potential conflict of interest was reported by the authors.

Funding

This work was supported by the National Natural Science Foundation of China under Grant No. 72174118 and Shanghai Philosophy and Social Science Project under Grant No. 2021ESH002.

Author contributions

Junyao Zheng: Conceptualization; Data curation; Investigation; Software; Visualization; Writing - original draft; Writing - review & editing. Yongbo Lu: Data curation; Investigation; Visualization; Writing - original draft; Writing - review & editing. Wenjie Li: Data curation; Investigation; Validation. Bin Zhu: Data curation; Writing - review & editing. Fan Yang: Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Software; Supervision; Writing - review & editing. Jie Shen: Supervision; Data curation; Writing - review & editing.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.

Ethics and other permissions

Not applicable.

References

Author notes