Performance of a rule-based semi-automated method to optimize chart abstraction for surveillance imaging among patients treated for non-small cell lung cancer.

BACKGROUND: We aim to develop and test performance of a semi-automated method (computerized query combined with manual review) for chart abstraction in the identification and characterization of surveillance radiology imaging for post-treatment non-small cell lung cancer patients. METHODS: A gold standard dataset consisting of 3011 radiology reports from 361 lung cancer patients treated at the Veterans Health Administration from 2008 to 2016 was manually created by an abstractor coding image type, image indication, and image findings. Computerized queries using a text search tool were performed to code reports. The primary endpoint of query performance was evaluated by sensitivity, positive predictive value (PPV), and F1 score. The secondary endpoint of efficiency compared semi-automated abstraction time to manual abstraction time using a separate dataset and the Wilcoxon rank-sum test. RESULTS: Query for image type demonstrated the highest sensitivity of 85%, PPV 95%, and F1 score 0.90. Query for image indication demonstrated sensitivity 72%, PPV 70%, and F1 score 0.71. The image findings queries ranged from sensitivity 75-85%, PPV 23-25%, and F1 score 0.36-0.37. Semi-automated abstraction with our best performing query (image type) improved abstraction times by 68% per patient compared to manual abstraction alone (from median 21.5 min (interquartile range 16.0) to 6.9 min (interquartile range 9.5), p < 0.005). CONCLUSIONS: Semi-automated abstraction using the best performing query of image type improved abstraction efficiency while preserving data accuracy. The computerized query acts as a pre-processing tool for manual abstraction by restricting effort to relevant images. Determining image indication and findings requires the addition of manual review for a semi-automatic abstraction approach in order to ensure data accuracy.

Thematic Analysis of Hospice Mentions in the Health Records of Veterans with Advanced Kidney Disease.

BACKGROUND: Patients with advanced kidney disease are less likely than many patients with other types of serious illness to enroll in hospice. Little is known about real-world clinical decision-making related to hospice for members of this population. METHODS: We used a text search tool to conduct a thematic analysis of documentation pertaining to hospice in the electronic medical record system of the Department of Veterans Affairs, for a national sample of 1000 patients with advanced kidney disease between 2004 and 2014 who were followed until October 8, 2019. RESULTS: Three dominant themes emerged from our qualitative analysis of the electronic medical records of 340 cohort members with notes containing hospice mentions: (1) hospice and usual care as antithetical care models: clinicians appeared to perceive a sharp demarcation between services that could be provided under hospice versus usual care and were often uncertain about hospice eligibility criteria. This could shape decision-making about hospice and dialysis and made it hard to individualize care; (2) hospice as a last resort: patients often were referred to hospice late in the course of illness and did not so much choose hospice as accept these services after all treatment options had been exhausted; and (3) care complexity: patients' complex care needs at the time of hospice referral could complicate transitions to hospice, stretch the limits of home hospice, and promote continued reliance on the acute care system. CONCLUSIONS: Our findings underscore the need to improve transitions to hospice for patients with advanced kidney disease as they approach the end of life.

The Feasibility of Using Large-Scale Text Mining to Detect Adverse Childhood Experiences in a VA-Treated Population.

Free text in electronic health records resists large-scale analysis. Text records facts of interest not found in encoded data, and text mining enables their retrieval and quantification. The U.S. Department of Veterans Affairs (VA) clinical data repository affords an opportunity to apply text-mining methodology to study clinical questions in large populations. To assess the feasibility of text mining, investigation of the relationship between exposure to adverse childhood experiences (ACEs) and recorded diagnoses was conducted among all VA-treated Gulf war veterans, utilizing all progress notes recorded from 2000-2011. Text processing extracted ACE exposures recorded among 44.7 million clinical notes belonging to 243,973 veterans. The relationship of ACE exposure to adult illnesses was analyzed using logistic regression. Bias considerations were assessed. ACE score was strongly associated with suicide attempts and serious mental disorders (ORs = 1.84 to 1.97), and less so with behaviorally mediated and somatic conditions (ORs = 1.02 to 1.36) per unit. Bias adjustments did not remove persistent associations between ACE score and most illnesses. Text mining to detect ACE exposure in a large population was feasible. Analysis of the relationship between ACE score and adult health conditions yielded patterns of association consistent with prior research.

Implications of Cross-System Use Among US Veterans With Advanced Kidney Disease in the Era of the MISSION Act: A Qualitative Study of Health Care Records.

Importance: Since 2014, when Congress passed the Veterans Access Choice and Accountability (Choice) Act (replaced in 2018 with the more comprehensive Maintaining Internal Systems and Strengthening Integrated Outside Networks [MISSION] Act), the Department of Veterans Affairs (VA) has been paying for US veterans to receive increasing amounts of care in the private sector (non-VA care or VA community care). However, little is known about the implications of these legislative changes for the VA system. Objective: To describe the implications for the VA system of recent increases in VA-financed non-VA care. Design, Setting, and Participants: This qualitative study was a thematic analysis of documentation in the electronic health records (EHRs) of a random sample of US veterans with advanced kidney disease between June 6, 2019, and February 5, 2021. Exposures: Mentions of community care in participant EHRs. Main Outcomes and Measures: Dominant themes pertaining to VA-financed non-VA care. Results: Among 1000 study participants, the mean (SD) age was 73.8 (11.4) years, and 957 participants (95.7%) were male. Three interrelated themes pertaining to VA-financed non-VA care emerged from qualitative analysis of documentation in cohort member EHRs: (1) VA as mothership, which describes extensive care coordination by VA staff members and clinicians to facilitate care outside the VA and the tendency of veterans and their non-VA clinicians to rely on the VA to fill gaps in this care; (2) hidden work of veterans, which describes the efforts of veterans and their family members to navigate the referral process, and to serve as intermediaries between VA and non-VA clinicians; and (3) strain on the VA system, which describes a challenging referral process and the ways in which cross-system care has stretched the traditional roles of VA staff and clinicians and interfered with VA care processes. Conclusions and Relevance: The findings of this qualitative study describing VA-financed non-VA care for veterans with advanced kidney disease spotlight the substantial challenges of cross-system use and the strain placed on the VA system, VA staff and clinicians, and veterans and their families in recent years. These difficult-to-measure consequences of cross-system care should be considered when budgeting, evaluating, and planning the provision of VA-financed non-VA care in the private sector.

Complexity and Challenges of the Clinical Diagnosis and Management of Long COVID.

Importance: There is increasing recognition of the long-term health effects of SARS-CoV-2 infection (sometimes called long COVID). However, little is yet known about the clinical diagnosis and management of long COVID within health systems. Objective: To describe dominant themes pertaining to the clinical diagnosis and management of long COVID in the electronic health records (EHRs) of patients with a diagnostic code for this condition (International Statistical Classification of Diseases and Related Health Problems, Tenth Revision [ICD-10] code U09.9). Design, Setting, and Participants: This qualitative analysis used data from EHRs of a national random sample of 200 patients receiving care in the Department of Veterans Affairs (VA) with documentation of a positive result on a polymerase chain reaction (PCR) test for SARS-CoV-2 between February 27, 2020, and December 31, 2021, and an ICD-10 diagnostic code for long COVID between October 1, 2021, when the code was implemented, and March 1, 2022. Data were analyzed from February 5 to May 31, 2022. Main Outcomes and Measures: A text word search and qualitative analysis of patients' VA-wide EHRs was performed to identify dominant themes pertaining to the clinical diagnosis and management of long COVID. Results: In this qualitative analysis of documentation in the VA-wide EHR, the mean (SD) age of the 200 sampled patients at the time of their first positive PCR test result for SARS-CoV-2 in VA records was 60 (14.5) years. The sample included 173 (86.5%) men; 45 individuals (22.5%) were identified as Black and 136 individuals (68.0%) were identified as White. In qualitative analysis of documentation pertaining to long COVID in patients' EHRs 2 dominant themes were identified: (1) clinical uncertainty, in that it was often unclear whether particular symptoms could be attributed to long COVID, given the medical complexity and functional limitations of many patients and absence of specific markers for this condition, which could lead to ongoing monitoring, diagnostic testing, and specialist referral; and (2) care fragmentation, describing how post-COVID-19 care processes were often siloed from and poorly coordinated with other aspects of care and could be burdensome to patients. Conclusions and Relevance: This qualitative study of documentation in the VA EHR highlights the complexity of diagnosing long COVID in clinical settings and the challenges of caring for patients who have or are suspected of having this condition.

Thematic Analysis of the Health Records of a National Sample of US Veterans With Advanced Kidney Disease Evaluated for Transplant.

Importance: To be considered for a kidney transplant, patients with advanced kidney disease must participate in a formal evaluation and selection process. Little is known about how this process proceeds in real-world clinical settings. Objective: To characterize the transplant evaluation process among a representative national sample of US veterans with advanced kidney disease who were referred to a kidney transplant center. Design, Setting, and Participants: This qualitative study was a thematic analysis of clinician notes in the electronic health records of US veterans referred for kidney transplant evaluation. In a random sample of 4000 patients with advanced kidney disease between January 1, 2004, and December 31, 2014, cared for in the US Department of Veterans Affairs (VA) health care system, there were 211 patients who were referred to a transplant center during the follow-up period. This group was included in the qualitative analysis and was followed up until their date of death or the end of the follow-up period on October 8, 2019. Main Outcomes and Measures: Dominant themes pertaining to the kidney transplant evaluation and selection process identified through thematic analysis. Results: Among 211 study patients, the mean (SD) age was 57.9 (9.5) years, and 202 patients (95.7%) were male. The following 4 dominant themes regarding the transplant evaluation process emerged: (1) far-reaching and inflexible medical evaluation, in which patients were expected to complete an extensive evaluation that could have substantial physical and emotional consequences, made little accommodation for their personal values and needs, and impacted other aspects of their care; (2) psychosocial valuation, in which the psychosocial component of the transplant assessment could be subjective and intrusive and could place substantial demands on patients' family members; (3) surveillance over compliance, in which the patients' ability and willingness to follow medical recommendations was an important criterion for transplant candidacy and their adherence to a wide range of recommendations and treatments was closely monitored; and (4) disempowerment and lack of transparency, in which patients and their local clinicians were often unsure about what to expect during the evaluation process or about the rationale for selection decisions. For the evaluation process to proceed, local clinicians had to follow transplant center requirements even when they believed the requirements did not align with best practices or the patients' needs. Conclusions and Relevance: In this qualitative study of US veterans with advanced kidney disease evaluated for transplant, clinician documentation in the medical record indicated that, to be considered for a kidney transplant, patients were required to participate in a rigid, demanding, and opaque evaluation and selection process over which they and their local clinicians had little control. These findings highlight the need for a more evidence-based, individualized, and collaborative approach to kidney transplant evaluation.

Care Practices for Patients With Advanced Kidney Disease Who Forgo Maintenance Dialysis.

Importance: Although maintenance dialysis is a treatment choice with potential benefits and harms, little is known about care practices for patients with advanced chronic kidney disease who forgo this treatment. Objective: To describe how decisions not to start dialysis unfold in the clinical setting. Design, Setting, and Participants: A qualitative study was performed of documentation in the electronic medical records of 851 adults receiving care from the US Veterans Health Administration between January 1, 2000, and October 1, 2011, who had chosen not to start dialysis. Qualitative analysis was performed between March 1, 2017, and April 1, 2018. Main Outcomes and Measures: Dominant themes that emerged from clinician documentation of clinical events and health care interactions between patients, family members, and clinicians relevant to the decision to forgo dialysis. Results: In the cohort of 851 patients (842 men and 9 women; mean [SD] age, 75.0 [10.3] years), 567 (66.6%) were white. Three major dynamics relevant to understanding how decisions to forgo dialysis unfolded were identified. The first dynamic was that of dialysis as the norm: when patients expressed a desire to forgo dialysis, it was unusual for clinicians to readily accept patients' decisions. Clinicians tended to repeatedly question this preference over time, deliberated about patients' competency to make this decision, used a variety of strategies to encourage patients to initiate dialysis, and prepared for patients to change their minds and start dialysis. The second dynamic arose when patients were not candidates for dialysis: clinicians viewed particular patients as not candidates or appropriate for dialysis, usually on the basis of specific characteristics and/or expected prognosis, rather than after consideration of patients' goals and values. When clinicians decided patients were not candidates for dialysis, there seemed to be little room for uncertainty in these decisions. The third dynamic occurred when clinicians believed they had little to offer patients beyond dialysis: when it was clear that patients would not be starting dialysis, nephrologists often signed off from their care and had few recommendations other than referral to hospice care. Conclusions and Relevance: These findings describe an all-or-nothing approach to caring for patients with advanced chronic kidney disease in which initiation of dialysis served as a powerful default option with few perceived alternatives. Stronger efforts are needed to develop a more patient-centered approach to caring for patients with advanced chronic kidney disease that is capable of proactively supporting those who do not wish to start dialysis.

Decisions about Renal Replacement Therapy in Patients with Advanced Kidney Disease in the US Department of Veterans Affairs, 2000-2011.

BACKGROUND AND OBJECTIVES: It is not known what proportion of United States patients with advanced CKD go on to receive RRT. In other developed countries, receipt of RRT is highly age dependent and the exception rather than the rule at older ages. DESIGN, SETTING, PARTICIPANTS, & MEASUREMENTS: We conducted a retrospective study of a national cohort of 28,568 adults who were receiving care within the US Department of Veteran Affairs and had a sustained eGFR <15 ml/min per 1.73 m2 between January 1, 2000 to December 31, 2009. We used linked administrative data from the US Renal Data System, US Department of Veteran Affairs, and Medicare to identify cohort members who received RRT during follow-up through October 1, 2011 (n=19,165). For a random 25% sample of the remaining 9403 patients, we performed an in-depth review of their VA-wide electronic medical records to determine the treatment status of their CKD. RESULTS: Two thirds (67.1%) of cohort members received RRT on the basis of administrative data. On the basis of the results of chart review, we estimate that an additional 7.5% (95% confidence interval, 7.2% to 7.8%) of cohort members had, in fact, received dialysis, that 10.9% (95% confidence interval, 10.6% to 11.3%) were preparing for and/or discussing dialysis but had not started dialysis at most recent follow-up, and that a decision had been made not to pursue dialysis in 14.5% (95% confidence interval, 14.1% to 14.9%). The percentage of cohort members who received or were preparing to receive RRT ranged from 96.2% (95% confidence interval, 94.4% to 97.4%) for those <45 years old to 53.3% (95% confidence interval, 50.7% to 55.9%) for those aged ≥85 years old. Results were similar after stratification by tertile of Gagne comorbidity score. CONCLUSIONS: In this large United States cohort of patients with advanced CKD, the majority received or were preparing to receive RRT. This was true even among the oldest patients with the highest burden of comorbidity.

Initial Steps toward Validating and Measuring the Quality of Computerized Provider Documentation.

BACKGROUND: Concerns exist about the quality of electronic health care documentation. Prior studies have focused on physicians. This investigation studied document quality perceptions of practitioners (including physicians), nurses and administrative staff. METHODS: An instrument developed from staff interviews and literature sources was administered to 110 practitioners, nurses and administrative staff. Short, long and original versions of records were rated. RESULTS: Length transformation did not affect quality ratings. On several scales practitioners rated notes less favorably than administrators or nurses. The original source document was associated with the quality rating, as was tf·idf, a relevance statistic computed from document text. Tf·idf was strongly associated with practitioner quality ratings. CONCLUSION: Document quality estimates were not sensitive to modifying redundancy in documents. Some perceptions of quality differ by role. Intrinsic document properties are associated with staff judgments of document quality. For practitioners, the tf·idf statistic was strongly associated with the quality dimensions evaluated.