Full text

Turn on search term navigation

Introduction

Malnutrition, or more specifically undernutrition, occurs when good health is limited by adequate consumption of calories and essential nutrients [1, 2]. Multicenter and international studies estimate that 20–60% of hospitalized patients are malnourished [2–6]. Malnutrition is associated with poor clinical outcomes, such as impaired wound healing [7–9], a longer length of stay [10–12], morbidity, and increased mortality [13, 14]. Additionally, providing adequate care for malnourished patients requires resource-intensive interventions and specialized nutrition support professionals, both of which contribute to the cost of caring for these patients [15, 16].

Despite the increased costs associated with the provision of care, hospitals struggle to capture appropriate reimbursement for malnourished patients due to both under-documentation and improper documentation of a malnutrition diagnosis [17–19]. Surgical patients, especially oncologic patients, have an exceptionally high prevalence of malnutrition, upwards of 60% [20]. Unfortunately, when these patients’ nutrition status is not addressed by licensed independent practitioners (LIPs), their nutrition interventions can be delayed or withheld [21]. While proper documentation of malnutrition in hospitalized patients by LIPs has been shown to positively affect hospital reimbursement, many hospitals utilize registered dietician nutritionists (RDNs) to screen and document these findings [9, 22, 23]. Unfortunately, thorough documentation does not directly impact reimbursement, leading to potential undercharging for a patient’s medical complexity [23, 24]. In 2012, the Academy of Nutrition and Dietetics and the American Society for Parenteral and Enteral Nutrition developed a consensus statement on the appropriate criteria to accurately identify malnutrition in adults [2]. Methods to assess patients for malnutrition include the Subjective Global Assessment (SGA) scoring system and the Nutrition-Focused Physical Exam (NFPE), both of which have utility in clinical settings [25–27].

The Medicare Severity-Diagnostic Related Group (MS-DRG) classification established by the Centers for Medicare & Medicaid Services (CMS) is determined by the principal diagnosis. Hospital reimbursement rates are based on the average cost required to care for patients with diagnoses of that MS-DRG. This reimbursement value increases depending on major complications and comorbidities (MCCs) or complications and comorbidities (CCs) associated with the care of that patient. A secondary diagnosis of malnutrition, depending on its severity and corresponding International Classification of Diseases, 10^th revision code (ICD-10), can qualify as both an MCC or a CC and can thus increase reimbursement. Additionally, interdisciplinary communication between LIPs and RDNs can improve accurate reimbursement for malnourished hospitalized patients by improving accurate documentation [24, 28–31].

Studies have shown that the implementation of a structured assessment for nutrition status in hospitalized patients can improve the detection and documentation of malnutrition, with a resulting positive impact on reimbursement [32]. Most studies evaluating malnutrition diagnosing and hospital reimbursement employ National Risk Screening (NRS) or SGA screening systems for nutrition status assessment. Unfortunately, few studies following this design have been conducted in the US on surgical patients [29, 30, 31]. Our study aimed to evaluate accurate documentation of nutrition status between RDNs and LIPs before and after the implementation of a dietitian-led NFPE intervention at an academic medical center in the southeastern US.

Methods

Due to the retrospective nature of this study, with no more than minimal harm to the study subjects, a waiver of informed consent was approved by our Institutional Review Board; IRB approval number 58780. Data were anonymized before analysis.

Intervention

Malnutrition scoring systems such as the NFPE were introduced to providers in April 2017. Registered dieticians provided education to physicians and advanced practice providers regarding best practices for documentation and diagnosis of malnutrition. Didactic sessions included demonstrations on how to use the NFPE, interactive case presentations, and hands-on learning of where to appropriately document this diagnosis in the medical chart.

Study population.

As part of a retrospective study, the electronic medical records of surgical patients from the University of Kentucky were queried from October 1, 2016, through January 31, 2018. Using ICD-10 codes related to malnutrition, patient malnutrition status of mild (E44.1), moderate (E44.0), severe (E43), and unspecified (E46) protein-calorie malnutrition were identified. All patients over the age of 18, admitted to a surgical service (vascular surgery, general surgery, colorectal surgery, surgical oncology) and evaluated by an RDN, were included in our analysis.

Patients were grouped based on their relationship to the intervention completed in April 2017. Therefore, patients from October 1, 2016, through March 31, 2017, were considered the pre-intervention cohort. Patients that were seen between April 1, 2017, and July 31, 2017, the three months following the intervention, were considered a transitional cohort. The transitional cohort was excluded from the final cost analysis. Finally, patients from August 1, 2017, through January 31, 2018, were considered the post-intervention patient cohort.

Outcome variables.

Following the data query, patient medical records were analyzed to determine the congruency of the documentation between the registered dietician and physician diagnoses. Patient cohorts were analyzed for both pre-intervention and post-intervention, broken out by surgical service line.

Using Medicare and Medicaid weighted DRG multipliers, estimated reimbursement outcomes attributed to malnutrition documentation were calculated. Reimbursements were followed at 6 months, 12 months, 18 months, and 24 months post-intervention to identify the sustainability of the intervention.

Statistical analysis.

Standard statistical analysis was applied to this dataset to provide the best depiction of data trends. Paired T-tests were utilized to compare pre-implementation and post-implementation averages. Additionally, non-linear generic method of moments (GMM) parameters were estimated for the study timeframe to determine the correlation between values. All statistics were done in Statistical Analysis System (SAS) statistical modeling software.

Results

Study population

In total, 528 patients were included for analysis. The pre-intervention cohort consisted of 194 patients, while the post-intervention group included 334 surgical patients. Transitional cohort patients were not included in the analysis. In both the pre-intervention and post-intervention groups, there were no statistical differences between demographics, length of stay, insurance status, age, cancer diagnosis, or ventilation rates (Table 1).

[Figure omitted. See PDF.]

Diagnosis documentation correlations.

Before the educational, interdisciplinary intervention, 8.4% of patients had congruent documentation between dieticians and providers. Following the intervention, 46.31% of documentation was congruent (p<0.001) (Table 2). When comparing patient time cohorts, before the intervention there is a monthly increase of about 2.35% in agreement. Following the educational intervention, the correlation between RDN and LIP documentation was about 13%. The correlation slope began to decrease by 1.45% starting at 11 weeks post-intervention. Despite the decreasing slope, the positive correlation of 0.8% persists through 30 weeks post-intervention (Table 3). No post-intervention washout phase or drift-back to previous documentation incongruencies was seen in the 30 weeks post-intervention (Fig 1).

[Figure omitted. See PDF.]

Cost.

Before the intervention period, 26 visits were estimated to have malnutrition diagnosis codes correctly applied to malnourished patients, resulting in a total of $278,566 additional reimbursement directly attributed to appropriate documentation of malnutrition based on Medicaid and Medicare reimbursement rates. During the transition phase, $204,594 of reimbursement related to 25 malnutrition-coded visits was estimated. Post-intervention, 68 visits were estimated to have a malnutrition diagnosis affected DRG, resulting in $571,281 of additional reimbursement. Follow-up estimates were completed at 6-month intervals through 24 months following the intervention. The estimated reimbursements related to malnutrition diagnosis codes rose from $669,129 at 6 months to $705,109, $704,102, and $835,634 at 12 months, 18 months, and 24 months, respectively (Table 4).

[Figure omitted. See PDF.]

Discussion

Malnutrition is associated with poor outcomes in surgical patients. While commonly documented by RDNs, there is frequently incongruent documentation by LIPs resulting in improper reimbursement for malnourished patients. Our study identified a successful quality improvement initiative led by RDNs to improve documentation of malnutrition by LIPs, which resulted in improved hospital reimbursement for malnourished patients.

Before our intervention, 8.4% of patients carried a billable diagnosis of malnutrition. This is similar to published rates of 0.6–18.0% based on population studies at discharge by Tobert et. al. [33], which identified a total of 5.8 million patient discharges over 2 years from 105 institutions with a reported malnutrition diagnosis rate of 4%. Hospital systems with improved diagnosis rates were associated with higher hospital rankings and improved patient satisfaction. The lack of LIP diagnosis, across specialties, raises concern regarding the identification of at-risk patients, given the known poor outcomes associated with malnourished patients [10]. Our study is the first to identify a sustainable intervention to improve the documentation and identification of malnourished surgical patients by LIPs. Following the intervention, 46.3% of patients had congruent documentation between the RDN diagnosis and LIP diagnosis.

Our intervention centered on an RDN-led educational campaign to educate LIPs on the utilization of the NFPE and associated diagnosis documentation. Studies have shown that the implementation of a structured assessment for nutrition status in hospitalized patients can improve the detection and documentation of malnutrition, with a resulting positive impact on reimbursement. In one such German study of 1244 patients undergoing elective surgery [34], the National Risk Screening (NRS) scoring system was used to screen patients’ nutritional status, and an appropriate ICD-10 code was assigned to each patient based on their NRS score. A simulation was run to assess the economic impact of proper coding based on the German-DRG system, and the authors found that the generated profit for each patient at risk for malnutrition was €263.96. In another similar German study of 541 patients in a gastroenterology ward, the SGA scoring system was used for screening and economic analysis of proper coding based on the ICD-10/German-DRG system which found a reimbursement increase of €360 per malnourished patient [35]. A Portuguese study of 129 patients in an internal medicine ward, using both the NRS and SGA scoring systems, also concluded that proper coding resulted in an overall increase in hospital reimbursement [36]. Other studies demonstrating shortfalls in hospital reimbursement due to poor documentation and coding of malnutrition also exist [17, 37]. Our study reveals a 3-times increase in reimbursement from the pre-intervention phase of $278,566 to the 24-month follow-up of $835,634 across all general surgery subspecialties.

Healthcare costs related to malnutrition are significant. Direct and indirect costs related to malnutrition are cited as up to 49 billion dollars; however, even these estimates are related to coded diagnoses rather than actual numbers of malnourished patients [38]. The cost of care for malnourished patients is increased with a longer length of hospital stay, surgical site infections, and in-hospital mortality [16, 29, 39]. In Canada, the Malnutrition Task Force utilized the SGA to identify sources of increased hospital costs in malnutrition. In 956 patients, hospital length of stay for both medical and surgical patients was increased by 23%, and 31%, respectively [16]. This corresponded to a 31% and 34% increased cost of hospital stay, respectively, when compared to well-nourished patients. Lee et. al. found surgical patients presenting with small bowel obstructions had a 4.2 times increased rate of in-hospital mortality if they also suffered severe protein-calorie malnutrition [21]. This study provides an intervention to improve the LIP identification of malnourished patients using the NFPE.

Physician awareness regarding malnutrition indicators and documentation is imperative to optimal patient care and appropriate reimbursement. Increased reimbursement funds could provide opportunities for funding pre-habilitation programs and other intensive nutrition programs to improve patient outcomes. While this study identifies a single application of multidisciplinary involvement to improve accurate documentation, hospital billing, and patient care, interdisciplinary collaboration benefits all vulnerable patient populations and should be pursued in all aspects of patient care. Additionally, similar multidisciplinary approaches for accurate documentation may aid in billing for other patient comorbidities, which may require significant healthcare resources.

Limitations

The limitations of this study are related to the estimated reimbursement rates based on Medicare and Medicaid DRG data. Institutional cost data was not available for analysis due to the inability to identify patients who were not coded by a diagnosis code by LIPs in our financial system. However, Medicare and Medicaid DRG reimbursement is a widely accepted estimation of costs and hospital financial data. While clinical outcomes such as length of stay, adverse outcomes, and mortality were not assessed in this quality improvement project, further research is needed to determine if accurate documentation has a meaningful impact on these important clinical metrics.

Conclusions

Overall, our study details the outcomes of an RDN-led educational quality improvement initiative to improve diagnosis ICD coding for malnourished patients. Our quality improvement initiative resulted in a sustainable 3-fold increase in reimbursement for malnourished patients and improved congruency between RDN and LIP documentation from 8.4% to 46.3%. Further quality improvement initiatives are needed to improve outcomes for malnourished patients and reduce overall hospital costs for this vulnerable population. Identification of malnourished patients is the first step to improving outcomes via targeted quality improvement initiatives. Therefore, further initiatives to improve interdisciplinary engagement should be explored to improve high-quality healthcare in surgical patients. Our project may inform other academic medical centers on how to proceed with quality improvement initiatives on similar patient outcomes topics.

Supporting information

S1 Dataset.

https://doi.org/10.1371/journal.pone.0287124.s001

(ZIP)

Citation: Levy BE, Castle JT, Wilt WS, Fedder K, Riser J, Burke ED, et al. (2023) Improving physician documentation for malnutrition: A sustainable quality improvement initiative. PLoS ONE 18(8): e0287124. https://doi.org/10.1371/journal.pone.0287124

About the Authors:

Brittany E. Levy