Risk identification and prediction of complaints and misconduct against health practitioners: a scoping review

Abstract Identifying the risk and predicting complaints and misconduct against health practitioners are essential for healthcare regulators to implement early interventions and develop long-term prevention strategies to improve professional practice and enhance patient safety. This scoping review aims to map out existing literature on the risk identification and prediction of complaints and misconduct against health practitioners. This scoping review followed Arksey and O’Malley’s five-stage methodological framework. A comprehensive literature search was conducted on MEDLINE, EMBASE, and CINAHL databases and finished on the same day (6 September 2021). Articles meeting the eligibility criteria were charted and descriptively analysed through a narrative analysis method. The initial search generated 5473 articles. After the identification, screening, and inclusion process, 81 eligible studies were included for data charting. Three key themes were reported: methods used for identifying risk factors and predictors of the complaints and misconduct, synthesis of identified risk factors and predictors in eligible studies, and predictive tools developed for complaints and misconduct against health practitioners. The findings reveal that risk identification and prediction of complaints and misconduct are complex issues influenced by multiple factors, exhibiting non-linear patterns and being context specific. Further efforts are needed to understand the characteristics and interactions of risk factors, develop systematic risk prediction tools, and facilitate the application in the regulatory environment.


Introduction
In recent years, health regulators have shown an increasing inclination towards proactive and data-driven approaches to oversight [1].Complaints and misconduct are recognized as valuable resources for understanding patient dissatisfaction and harm, as well as detecting problematic practitioner practices [2,3].Regulatory agencies worldwide consider complaints and misconduct as crucial components of their risk management frameworks.For instance, in the USA, the Vanderbilt Center for Patient and Professional Advocacy (CPPA) has developed a Patient Advocacy Reporting System (PARS) complaint programme in collaboration with hospitals and medical groups to identify physicians at high risk of complaints effectively [4].Similarly, national schemes have been implemented in Australia and New Zealand to regulate health professions, with a particular focus on complaints and disciplinary data [5].These developments were noted as offering 'opportunities for researchers, policymakers and regulators to move towards evidence-based regulation' [6].
In related fields, a growing number of studies have explored complaints and misconduct against health practitioners.Numerous investigations have revealed that the occurrence of such complaints and misconduct is not entirely random, with a small proportion of practitioners accounting for up to 50% of claims [7][8][9].Certain research has identified specific characteristics associated with high-risk health practitioners, including demographics such as gender, age, and speciality [10,11].Moreover, beyond these "personal" or "human" factors, practice or system characteristics, such as workload [12] and practice setting [11], have emerged as discernible indicators of medico-legal matters.Despite the growing body of research in this field, there is a lack of reviews that comprehensively examine the risk of complaints and misconduct against health practitioners, especially when viewed through a predictive and prospective lens.
This study adopts a broad perspective with the goal of providing a comprehensive understanding of existing literature on the risk and prediction of complaints and misconduct against health practitioners.To achieve this, a broad scoping review methodology was employed due to its advantages in mapping a wide range of available evidence, identifying key characteristics or factors related to a concept, and conveying the breadth and depth of a field [13][14][15].This study seeks to address the following three specific questions: (i) what methods have been used to identify risk factors and predictors of complaints and misconduct against health practitioners?(ii) What are the patterns and taxonomy of risk factors and predictors of complaints and misconduct against health practitioners?(iii) What predictive models or tools have been developed to identify risks for complaints or misconduct against health practitioners prospectively?

Study design
This study utilized the well-established five-stage methodological framework proposed by Arksey and O'Malley [13], which is widely recognized and commonly used in scoping reviews.The five stages include (i) identifying the research question, (ii) identifying relevant studies, (iii) study selection, (iv) charting the data, and (v) collating, summarizing, and reporting the results.To ensure rationality and integrity, the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) extension for Scoping Reviews (PRISMA-ScR) checklist was also applied to report the findings [16].

Research question
The research question was formulated by following the Population, Concept, and Context framework [17].In this study, the Population encompasses all health practitioners.The Concept focuses on the identification and prediction of risks associated with complaints and misconduct.Within the scope of this study, the term "complaints and/or misconduct" is a broad term that encompasses various related concepts, including complaints, misconducts, disciplinary matters, malpractices, grievances, misbehaviours, medico-legal cases, medical errors, disputes, criminal cases, and claims that are against health practitioners.There are no geographic or contextual restrictions.Consequently, the following research question was posited: what is known from the existing literature about the identification and prediction of risks associated with complaints and misconduct against health practitioners?

Search strategy
The search strategy was developed over several research team meetings based on three sets of primary concepts: risk identification or prediction, health practitioners, and complaints and/or misconduct.A search of literature was conducted on 6 September 2021, using three databases: Medline via PubMed (1946 to the present), EMBASE via Ovid (1980 to the present), and CINHAL Plus via EBSCOhost (1937 to the present).All searches were subject to date and language limitations, with only English articles published since 2000 being included.Although the search was pre-planned in advance, no formal protocol was registered for this review.A full search strategy for each database is available in the supplementary file (Supplementary Appendix A).

Screening and selection process
The screening and selection process was conducted in three stages.Firstly, the search results were imported into Endnote software.Duplicates were removed using the software's built-in function, followed by further manual checking.Subsequently, the titles and abstracts of all imported studies were screened based on the inclusion criteria outlined in Table 1.Additionally, the reference lists of review papers were searched for additional relevant publications (i.e.backward snowballing) [13,18], ensuring comprehensive coverage of evidence.Finally, exclusion criteria were developed for full-text screening, and the final decision regarding the inclusion of the remaining articles was made based on the refined eligibility criteria.
To expedite the process, a rapid review approach "single screening" was adopted for abstract and full-text screening [19].This widely accepted rapid review approach offers a more efficient analysis and quicker generation of evidence portfolio.However, it does carry the risk of potentially missing relevant studies.Nevertheless, given the extensive number of studies identified through the searching and screening processes, any missing studies are expected to have minimal impact on the overall summarizing and dissemination of these research findings [20,21].
Throughout the searching, screening, and assessing process, the research team held regular meetings to discuss the eligible criteria, key concepts, and observations.One reviewer initially assessed each article, with any uncertainties resolved through discussions with two advisors who had expertise in pharmacy, organization management, and law until a consensus was reached.

Data extraction, analysis, and report
An Excel ® sheet was created to extract the following information: author(s), title, journal, year of publication, location, methodology, data source, year of the study conducted, involved professional(s), aims, key findings, and conclusions.A narrative synthesis analysis method recommended by Arksey and O'Malley (2005) was then performed, including numerical statistical summaries, textual commentaries, and tabular and graphical representations [22].

Search results and characteristics
The database search initially yielded 5473 articles.After removing 2212 duplicates, a total of 3261 articles were reviewed in title and abstract, and 292 were retained for full-text screening.Also included were 62 articles through screening the reference list of eight systematic reviews and traditional literature reviews.Through the full-text screening of 354 articles, 81 articles were finally included for data charting.The complete identification, screening, and inclusion process is illustrated in the PRISMA-ScR flowchart (Fig. 1).The supplementary file (Supplementary Appendix B) provides basic information on eligible studies.

Inclusion criteria
Exclusion criteria 1. Original research articles describing experimental or observational investigations or review articles using formal methods to explore the identification or prediction of the risk of complaints and misconduct against health practitioners.2. The primary focus of the study was on complaints, misconducts, disciplinary matters, malpractices, grievances, misbehaviours, medico-legal cases, medical errors, disputes, criminal cases, and claims against health practitioners (collectively referred to as "complaints and/or misconduct").3. Published in English since 2000.

Studies that focused on:
(i) health complaints that referred to patients' bodily disorders, diseases, symptoms, or distress; (ii) defensive medicine.2. The following types of prediction studies: (i) did not target health practitioners or only focused exclusively on veterinarians; (ii) focused solely on predicting specific outcomes of complaints or misconduct (such as payment or liability); (iii) only analysed patient or complainant-related risk factors.3. Publications in the form of commentary, editorials, letters, notes, news, communications, books, case reports, theoretical essays, conference abstracts, and oral presentations.4. Non-English studies or publications before 2000. 5.The completed text was not available through the university library service.

Methods used for identifying risk factors and predictors
The majority of the eligible studies (n = 76) adopt statistical analysis methods to identify risk factors or predictors of complaints and misconduct against health practitioners, encompassing 73 quantitative studies and 3 mixed-method studies.Specifically, 19 studies compared the characteristics and identified high-risk characteristics of complaints and misconduct cases by conducting significance tests such as the Chi-square test, Student's t-test, and Mann-Whitney test.a Review articles that collected data from more than one country.
Additionally, 57 studies further examine the risk factors associated with complaints or misconduct through regression models, such as Poisson regression [47], logistical regression, recurrent-event survival analysis, and generalized estimating equation.In five other eligible review articles, three employed traditional narrative review analysis, with two utilizing metaanalysis methods for analysis for the identification of risk factors.

Synthesis of identified risk factors and predictors
From the eligible 81 studies, three main categories of risk factors and predictors have been identified: health practitioners' factors, system and environment factors, and complaint/misconduct issues.The synthesises and detailed findings for each eligible study are provided in Supplementary Appendix C and D.
System and environment risk factors pertain to the factors related to the system, organization, or the broad environment in which the health practitioner practices.System factors included specific practice settings, practice size, facility type, and organization characteristics.One predominant system factor identified in the studies was the academic/university settings.Regarding environmental factors, several studies conducted in the USA [29,31,55,68], Australia [1,2,51,56], Denmark [30], and Canada [85] examined the relationship between geographic regions and the risk of complaints and misconduct.In addition, a Danish study found that General Practitioners serving lower-income patients were more likely to receive complaints or disciplinary board criticisms [50].
In terms of the issue of complaints and misconduct, four studies reported specific issues that pose a high risk.An Australian study found that complaints related to substance use, honesty, mental health, sexual boundaries, and medication use and supply were strongly recurrent [1].In a dental school setting, appointments had a high complaint rate, with communication being the second-highest complaint type overall [24].In the USA, one study revealed that plaintiffs overwhelmingly alleged diagnosis and procedural errors [43].Another American study identified dissatisfaction with treatment and accessibility as the two most common complaint types among high-risk otolaryngologists [54].
In 2019, an updated Predicted Risk of New Event for Health Practitioners (PRONE-HP) version was released, incorporating two additional predictors: practice location and complaint issues [1].

Statement of principal findings
The principal findings of this scoping review provide a comprehensive overview of the existing literature on risk identification and prediction related to complaints and misconduct against health practitioners.Although the included studies demonstrate a broad geographic and professional diversity, 60% were conducted in the USA and Australia, and 86% primarily focused on physicians.This observation suggests a need for more diverse and inclusive research encompassing a wider range of health practitioners in different contexts.
The identified risk factors and predictors fall into three categories: health practitioners' factors, system and environment factors, and issues of complaints/misconduct.Health practitioners' characteristics were the most frequently reported risk factors.Our analysis consistently found certain characteristics recurring as risk factors in more than three studies, such as older age, male gender, poor performance at school, complaints or misconduct history, longer in practice, and higher workloads.However, when investigating specific risk factors, it becomes evident that these factors are context specific and profession dependent.For example, studies vary in findings on age's impact on complaints and misconduct against health practitioners [1,4,76].Furthermore, different analytical perspectives reveal diverse patterns.Physician-level analysis shows higher operative volume heightens liability risk due to potential adverse events in every patient interaction, while institutionally, greater clinical volume correlates with lower mortality rates for some surgeries [49].This highlights the need for targeted and comprehensive studies to enhance the specific field across different settings and professions.
It is worth noting that environment and organization characteristics play an important role in risk assessment and prediction.While our review has included isolated findings mentioning these factors as potential risk factors for complaints and misconduct, few studies systematically examined them as distinct categories of risk factors, with only two exceptions [50,65].One repeatedly validated factor in this realm is the presence of academic/university settings.Factors contributing to this might involve treating patients with complex diagnoses, additional research, teaching and administration responsibilities, and complex organizational structures [55].
Another noteworthy finding is that the issue of complaints or misconduct itself has emerged as a risk factor, suggesting that specific issues can indicate different levels of risk.While this aspect was not extensively emphasized in eligible research, complaint issues have proven significant in identifying and predicting complaints and misconduct risks in practical predictive tools.Our findings reveal that two identified predicting tools incorporate complaint types as indicators-the CPPA PARS system profiles complaint types to reflect complaint distribution related to individuals [55].Conversely, the PRONE-HP score system provides a summary of the overall complaint-type distribution among health practitioners, revealing strong associations between complaints related to mental health, substance use, sexual boundaries, and honesty with recurrence [1].However, these models have not integrated external system or environmental factors, making it a potential area for further explorations.

Strengths and limitations
To our knowledge, this study is the first comprehensive review that synthesizes the issue of risk identification and prediction in the broad field of health practitioners.This expansive research scope enabled us to acquire a thorough understanding of all medico-legal matters related to complaints and misconduct, thereby showcasing a comprehensive landscape within the relevant field.The inclusion of all health practitioners allowed us to gain insights into current research patterns and identify sub-areas that require further exploration.Moreover, this study goes beyond the research progress but also elucidates the practical application, enriching the significant practical value of this research.
However, this study has a few limitations to consider when interpreting the results.Firstly, the language and publication date restrictions may have resulted in missing studies from non-English speaking countries as well as older studies, potentially limiting the scope of the review.Additionally, using a rapid review approach may have introduced some bias into the review process.However, the involvement of the two experts in all screening and assessment processes as much as possible helped to mitigate this potential bias to some extent.

Interpretation within the context of the wider literature
The findings of this study are in alignment with previous review articles regarding some category of risk factors associated with complaints or misconduct [10,11].A systematic review identified risk factors associated with malpractice claims and impaired performance in medical practitioners and categorized these factors as demographic and workplace related, encompassing practice-related elements such as training location, speciality as well as system, organization, and environmental factors related to clinical workload, clinical practice setting, and geographical location [10].Another focused mapping review synthesized the priori risk factors associated with medical practitioners, placing emphasis on the influence of "human factors" and system factors [11].This study expands the findings of existing reviews by highlighting the predictive function of complaints/misconduct issues, which have received little attention and were not explicitly suggested until 2019 [1].
In practice, most existing techniques used to analyse patient complaints are complex and pose challenges when implemented in frontline supervision.This is largely due to the lack of technical capacity among regulatory agencies [83].While the two tools developed so far have addressed this issue to some extent, they still come with some limitations, as both were primarily tailored for specific contexts, limiting their broader applicability.Furthermore, these models do not incorporate system and environment factors.Regulators may face complexity when trying to evaluate these factors across their entire caseload, making them unsuitable candidates for routine-use risk calculators.Given the importance of these contextual factors, future investigations should explore strategies to include them as variables in practical risk measurement [83].

Implications for policy, practice, and research
This study offers valuable insights for policymakers, educators, regulators, and practitioners, shedding light on the risk areas that need to be addressed, enabling the development of targeted regulatory and educational strategies.The findings also point towards several directions for future research.While robust evidence has been generated from quantitative analysis, we believe that incorporating qualitative approaches would be beneficial in gaining a deeper understanding of the nature, meaning, and mechanisms underlying high-risk and predictive characteristics of complaint and misconduct cases.In addition, the complexity observed in this study and previous research highlights the need to explore the interrelationships among these factors and the potential underlying mechanisms in different contexts.Furthermore, developing more comprehensive and user-friendly predictive tools that can convert research findings into a scoring system to guide interventions is essential.Testing the effectiveness of current prediction models in diverse health settings might also be likely to benefit from refining these predictive tools in practice applications.

Conclusion
This scoping review advances our understanding of the risk identification and prediction of complaints and misconduct against health practitioners.The findings highlight currently understudied areas, particularly related to qualitative analysis of risk factors, the indicative roles of complaint issues, and the development of practical tools.Furthermore, predicting the complaint risk is multifactorial, non-linear, and context specific.Future in-depth explorations in specific areas will significantly contribute to developing more proactive and data-driven regulations.

Figure 1
Figure 1 PRISMA flowchart for study selection.

Figure 2
Figure 2 Risk factors for predicting complaints and misconduct against health practitioners.Note: Risk factors in bold were identified in the literature more than three times.

Table 1 .
Inclusion and exclusion criteria.

Table 2 .
Quantitative description of identified publications.

Table 3 .
Predictive model and tools for complaints and misconduct analysis.