Clinical documentation integrity: Its role in health data integrity,patient safety and quality outcomes and its impact on clinical coding and health information management

Wide-ranging impact of CDI on various stakeholders

In their editorial “Recognising Complexity” Hemsley and Debono (2022) discussed the many influences on clinical documentation by multiple aspects of diversity across populations and cultures. Robinson et al. (2023) also drew attention to the multi-dimensional nature and multi-directional use of health information derived from clinical documentation, which can significantly impact patient safety, funding, research, health policy and planning. First-level stakeholders include patients, clinicians, CCs and HIMs and CDIS. Second-level or downstream stakeholders include those using administrative data, in particular the clinical codes and diagnosis-related groups (DRGs) that make up a large part of admitted patient datasets.

For patients and clinicians, the integrity of clinical documentation has been recognised as critical for quality and safety in patient care, yet also poorly understood and enacted. These concerns were reflected in a number of articles. For example, Chen et al. (2022) found that documentation of working diagnoses was highly variable among non-consultant grade clinicians, whereas McLachlan et al. (2023) reported that incomplete and inaccurate documentation of adverse drug reactions was common in the hospital under study. Sierla and Dylke (2023) also noted inconsistencies in how data on lymphoedema were gathered and understood, proving a barrier to sharing and comparing data among clinicians. Clinical documentation was even considered time-consuming and burdensome by some clinicians (Rowlands et al., 2022), whereas for traditional medicine practitioners in South Africa, lack of clinical documentation was associated with lack of experience and knowledge about what and how to document (Zhandire et al., 2021, 2023).

For CCs and HIMs, the integrity of clinical documentation is critical for accurate clinical coding. Kilkenny et al. (2022) reported variation in the reporting of comorbidities of stroke in administrative data suggesting the need to improve clinical documentation. Duke et al. (2022) examined coding-based definitions for sepsis and identified significant differences in sepsis classification in administrative data (e.g. use of explicit ICD10-AM diagnosis codes was poorly predictive of sepsis, with low sensitivity). Across Canada, Doktorchik et al. (2020) reported that clinical coding managers perceived data quality to be limited by incomplete and inconsistent chart documentation, with increasing expectations for data collection without appropriate allocation of resources. Jebraeily et al. (2023) found causes of clinical coding problems standing in the way of quality improvement fell into five groupings: policies, protocols and processes; individual factors; equipment and materials; working environment and management factors; and quality of clinical coding could be improved by hospital managers and health policymakers focusing on strategies and solutions that target these root causes. Hosseini et al. (2022) argued that data quality enhancement requires collaboration between physicians, nurses, managers and developers of software, suggesting a concept mapping approach to meet this objective. Kilkenny et al. (2023) recommended ongoing education for CCs to achieve high-quality coding and more accurate data. Campbell and Giadresco (2020) reviewed 39 articles on computer-assisted clinical coding (CAC) and found it improved both accuracy and quality of data. They argued that CAC would facilitate the development of clinical coding skills and knowledge, and advance the careers of clinical coding professionals.

The integrity of clinical documentation is also critical for DRG reporting and other requirements and optimisation of coding and DRG groups at this level has significant second-level implications, including financial. In their editorial on the role of health classifications in health information management, Shepheard and Groom (2020: 85) made the point that DRGs developed from the analysis of the clinical codes in administrative databases are one of the most high-profile uses of clinically coded data. Research by Alonso et al. (2020) clearly demonstrated how quality data resulted in more accurate payments, whereas Hay et al. (2020) also drew attention to the association between CDI and improved quality and patient safety outcomes as well as increased hospital reimbursement and funding.

For policy makers and researchers, the “use of coded data is extensive, robust and growing” (Riley et al., 2023a: 1). For example, to build a foundation for a clinical support system, Sierla and Dylke (2023) developed a core dataset comprised of required items for inclusion to digitise monitoring and reporting of upper limb lymphoedema. Palamuthusingam and Pascoe (2023) assessed concordance between comorbidities in the Australia and New Zealand Dialysis and Transplant Registry and state-based hospital admission datasets. They found agreement between the Registry and hospital datasets was variable, with prevalence of comorbidities being higher in the Registry. A study by Duke et al. (2022) also found the most frequently used coding method underpinning administrative data used for epidemiologic surveillance underestimated its prevalence. Hay et al. (2020) reported that only 55% of laboratory-confirmed episodes of Staphylococcus aureus bacteraemia were reflected in the coded data, suggesting that inaccuracies were most likely due to documentation issues. Kilkenny et al. (2022, 2023) examined the importance of accurate coded diagnostic data for epidemiological research into stroke and suggested that given the variation in reporting of comorbidities of stroke in administrative data, multiple sources of data may be necessary for research. While these datasets are critical, in many cases their users have little or no understanding of how the data are collected (Hemsley and Debono, 2022). Conversely, creators of clinical documentation may have limited appreciation of the wide-ranging impact of CDI and its coded derivatives, including for purposes of research and epidemiology. The potential dangers of researchers using data inappropriately for secondary purposes was highlighted by Riley et al. (2023b) who conducted a “fit for purpose” analysis of government health information assets. While the authors verified that the health information assets they examined had been appropriately used for secondary purposes, they nevertheless called for greater transparency in the secondary use of data, to safeguard public trust in the collection, storage and management of personal and sensitive information.

Strategies to address CDI

In this selected collection of articles, CDI strategies broadly targeted education, standardisation (forms, terminology, language), assistive technologies to support existing and emerging roles, (near) real-time workflow activities (CAC, disease surveillance), monitoring and feedback, embedding the coding/CDI workforce in clinical settings and highlighted the importance of effective relationships/communication, adequate resourcing and management support. Education or training was often recommended to improve health professionals’ understanding of what and how to document. However, clinicians were reported to be primarily motivated by patient care but considered poorly trained in the art of good clinical documentation (Rowlands et al., 2022). Recommendations indicated that clinician education could be extended beyond initial training, be ongoing and encompass clinical classifications and coding standards (Hosseini et al., 2022). Education for other stakeholders included targeted training for CCs (Doktorchik et al., 2020; Kilkenny et al., 2023), together with CDI audits to monitor, raise awareness and provide feedback (Alonso et al., 2020; McLachlan et al., 2023). The importance of building strong professional relationships and communication is imperative for collaborative working in all CDI activities, according to Doktorchik et al. (2020), Hosseini et al. (2022) and Pine et al. (2023).

Several authors (Chen et al., 2022; Hosseini et al., 2022; Payton et al., 2023; Rowlands et al., 2022) sought to understand clinician preference and decision-making in relation to assignment of diagnoses and documentation. Lack of knowledge and need for further education among junior clinicians was highlighted by Chen et al. (2022), and Hosseini et al. (2022) also recommended updated clinical training programmes, together with electronic health record (EHR) documentation guidelines and monitoring and feedback mechanisms. Payton et al. (2023) reported the need for balance between structured and narrative forms of documentation and that highly structured formats (a feature of EHRs) can limit usability. Several authors highlighted multifactor influences on CDI, including personal, systemic and environmental (Jebraeily et al., 2023; Lövestam et al., 2022; Rowlands et al., 2022).

The CDIS role in relation to clinician education is clearly aimed at addressing knowledge gaps through real-time interface with clinicians. This was identified by Pine et al. (2023) and also Hay et al. (2020), as a key CDI strategy, both foundational (part of clinical training programmes) and continuing. Hosseini et al.’s (2022) concept mapping method was used to identify the reasons why clinicians may be suboptimal documenters of clinical information. Workload and issues (such as late reporting of laboratory results) were also identified as problems, importantly (for the role of CDIS), lack of classification knowledge and lack of understanding of key clinical coding definitions such as principal diagnosis definition among clinicians. While CCs are well versed in the intricacies of health classifications and coding standards, they are also required to maintain currency with clinical terminology and clinical concepts. This can be challenging and can impact on the datasets relied upon by clinical programme managers. Kilkenny et al. (2023) described an education programme specifically for CCs to assist them to optimally interpret clinical documentation in relation to stroke patients. CCs are also charged with translating clinical documentation into coded form to be held in administrative datasets and made available to a wide range of stakeholders.

The significance of clinical coding as impacted by CDI has been alluded to in a number of articles in the VSI collection that describe issues experienced by specific clinical programmes where information was not available in a form that supported the programme. Strategies to address the limitations of administrative coded data and lack of specificity in support of CDI have been proposed by Duke et al. (2022), Payton et al. (2023) and Sierla and Dylke (2023). In their editorial on the emergence of health information in aged care, Loggie and Davis (2023) discussed needed improvements in clinical documentation related specifically to the aged care sector. Other recommended strategies to improve clinical documentation include the implementation of structured note templates, standardised terminology and the use of core datasets in diverse areas such as drug reactions (McLachlan et al., 2023), nutrition (Lövestam et al., 2020), stroke (Kilkenny et al., 2022), lymphoedema (Sierla and Dylke, 2023) and hernia surgery (Vetter and Kim, 2023).

Standardisation to address CDI-related issues (e.g. ambiguous terminology, inconsistent, incomplete and variability in diagnosis documentation) was recommended by Alonso et al. (2020) and Payton et al. (2023). Chen et al. (2022) described issues arising from documentation that is ambiguous when terms such as “probable,” “likely” and “query” are used. Doktorchik et al. (2020) and Jebraeily et al. (2023) both described incomplete and inconsistent chart documentation as barriers to coding quality. EHRs enable the use of standard terminologies such as SNOMED CT where clinical descriptors with increased specificity can be utilised compared to those that are available in classifications (e.g. ICD-10 and its Australian derivative, ICD-10-AM). The concept of specificity refers to the level of detail observable within data and is considered an “essential characteristic of high-quality data” (Roberts et al., 2023: 8). While specificity is recognised as important, it does not commonly feature in administrative data, often necessitating use of multiple data sources or data linkage for research and analytics (Duke et al., 2022; Kilkenny et al., 2022; Palamuthusingam and Pascoe, 2023; Roberts et al., 2023).

Several articles have described specific roles and positions engaged in CDI programmes. The roles of CCs, HIMs and CDIS inherently aim for greater specificity within clinical documentation to support patient care, clinical coding and other business activities. The CDIS role will support the interplay between clinical documentation, classifications and terminologies into the future. As Pine et al. (2023) described, CDIS and CDI programmes were originally established in the United States of America (USA) to support reimbursement activities and they continue to evolve. Their strengths are in building relationships and capabilities to enable timely adaptation to emerging CDI needs; COVID-19 surveillance was used as an example. While CDIS roles and CDI programmes now commonly feature in USA settings (Pine et al., 2023), similar roles are also evident in articles from Australia, the United Kingdom and Portugal. In these embedded roles, the CDIS concurrently review clinical documentation and provide near real-time (timely) feedback to the clinician. Several articles have exposed potential threats to existing roles and challenge future expectations of traditional roles. Identified CDI roles include clinical coding analysts/editors (Campbell and Giadresco, 2020), clinical coding quality coordinators (Doktorchik et al., 2020) and medical coders and auditing physicians (Alonso et al., 2020).

Impact of technology on CDI

Technology and data-intensive healthcare environments are changing the way clinical documentation is recorded and how it is used by clinicians and CCs (Pine et al., 2023). Increasing innovation and assistive technologies including EHR and electronic medical records (EMRs), CAC and EHR disease/public health surveillance have been described in several articles. While challenges exist, there is clear potential and interest in optimising data captured as electronic forms of clinical documentation; however, its integrity remains critical. Several articles considered the usability of EHR documentation from clinician perspectives (Lloyd et al., 2023; Payton et al., 2023). Lloyd et al. (2023) focused on situational factors impacting clinicians such as increasing complexity, work pressures, burnout and highlighting challenges identified by clinicians related to EMR such as usability, lack of intuitiveness, complexity, difficulties communicating with primary and other care sectors and time taken to perform clinical tasks.

Campbell and Gaidresco (2020: 5) defined CAC is as an “automated clinical coding process that assigns diagnoses and procedures from electronic sources of clinical documentation.” The literature examining CAC, predominantly from the USA, has reported improvements in clinical coding accuracy and quality compared to manual coding, although efficiency was influenced by workflows and maturity of information systems. In their review, automated clinical coding was perceived as a potential threat to existing clinical coding roles, while also representing career expansion opportunities as clinical coding editors and analysts (Campbell and Gaidresco, 2020). The historiography of the health information management profession by Robinson et al. (2023) similarly highlighted how change brings challenges and risks, and wider implications for education and practice.

While the emergence of EHRs has provided wonderful opportunities for improved patient care, such as legible clinical documentation reducing the rate of drug administration errors (McLachlan et al., 2023), it has also created challenges for data integrity. Campbell and Giadresco (2020) described issues of illegible writing within scanned records, abbreviations and symbols unable to be interpreted by electronic software and incomplete medical statements in patient records. Hosseini et al. (2022) also examined clinician factors impacting diagnosis documentation within EHR and paper-based records, concluding the need for documentation guidelines and to increase knowledge and awareness of clinical coding and the International Classification of Diseases, specifically within clinical training programmes.

Computer assistive technologies can be formal and informal, including Artificial Intelligence and Google as examples. Alonso et al. (2020) described the use of Google to decipher unfamiliar abbreviations/acronyms used in clinical documentation despite having standards and lists of accepted abbreviations available. Lack of information systems integration in both hybrid and electronic forms remains an ongoing threat to CDI. Timely and complete access to health records and key documents containing clinical documentation, including EHRs, are continuing concerns for clinicians and clinical coding (Alonso et al., 2020; Schwarz et al., 2023).