Abstract
Abstract
Background:
Prognostication plays a key role in palliative care (PC). It is critical for advance care planning, determining hospice eligibility, and communication. In contrast to subjective clinical prognostication, evidence-based prognostication (EBP) uses existing validated data to quantify prognosis; however, the extent to which PC providers use EBP is limited.
Objective:
The objective was to analyze documentation of EBP by PC providers in the absence of an inpatient consultation note template at a single academic medical center.
Methods:
We retrospectively evaluated prognostic documentation of inpatient PC consultations on oncology patients at a single academic hospital. Ratings of Eastern Cooperative Oncology Group (ECOG) Scale, Karnofsky Performance Scale (KPS), Palliative Performance Scale, and/or activities of daily living (ADLs) were considered documentation of functional status. PC-specific documentation of EBP included the Palliative Prognostic Index and/or Palliative Prognostic Score.
Results:
There were 412 inpatient PC consultations for oncology patients (2012–2013). Reasons for consultation included goals of care (n=108), symptom management (n=181), or both (n=123). In the absence of a note template, functional status was documented in 6% (n=24) of consultation notes, while no consultation notes contained EBP documentation of the Palliative Prognostic Index and Palliative Prognostic Score.
Conclusion:
This retrospective analysis conducted at a single academic medical center suggests poor documentation by PC providers of EBP in the absence of a consultation note template. Research and educational opportunities exist to evaluate barriers to EBP utilization and documentation by PC providers.
Introduction
F
Medical providers currently use two methods of prognostication: clinical prediction of survival and evidence-based prognostication (EBP). Clinical prediction of survival relies on a clinician's personal knowledge and experience to make subjective assessments regarding a patient's prognosis. 5 Unfortunately, clinical prediction of survival is often inaccurate.3,6 As a result, discussion regarding prognosis is often framed in broad time intervals, such as minutes to hours, hours to days, days to weeks, weeks to months, or months to years.3,6
The inaccuracy of clinical prediction of survival suggests alternative approaches are necessary. One alternative is the use of existing EBP, which has been demonstrated to be more accurate.7,8 PC-specific EBP includes the Palliative Prognostic Index and Palliative Prognostic Score. Assessments of functional status and identification of associated key signs, symptoms, or biological parameters are additional prognostic factors. 9 Despite the availability of EBP, utilization and documentation of EBP by PC providers may be challenging due to unfamiliarity with the data, limited access, and time constraints in completing calculations.
One recent retrospective analysis of 400 consecutive PC consultations at a single academic medical center demonstrated 72% of consultation notes included survival estimates. 3 However, survival estimates conducted were derived from clinical prediction of survival in broad time intervals (hours to days, days to weeks, etc.) and not by EBP. Furthermore, the authors encouraged further research in this area. 3 Consequently, our PC group evaluated current EBP documentation in an electronic medical record at our single academic medical center. In particular, we focused on documentation of PC-specific EBP and functional status in oncology patients with the goal to identify areas for growth and improvement.
Methods
Patient population and setting
IRB approval was obtained by the University of California, San Diego Human Research Protections Program. This study was completed at a single academic medical center with PC services provided at three hospitals with more than 500 acute care beds. The primary medical team requested new consultations from the inpatient PC team consisting of a physician, nurse practitioner, licensed clinical social worker, and pharmacist.
Patient data were collected for new inpatient consultations from January 2012 to January 2013 from a single electronic medical record. During this time there were more than a thousand requests for inpatient PC consultations on 685 unique patients. Study inclusion criteria included first-visit consultations in adult patients with cancer. For patients with multiple first-visit consultations from different hospitalizations, only the first-visit consultation note from the most recent hospitalization was included. The primary reasons for consultation were to assist with goals of care and/or symptom management.
Data collection
All members of the PC team wrote consultation notes. Patient demographics, functional status, and EBP were obtained from a single electronic medical record. Eastern Cooperative Oncology Group (ECOG) Scale, Karnofsky Performance Scale (KPS), Palliative Performance Scale, and activities of daily living (ADLs) were considered documentation of functional status.10–12 The Palliative Prognostic Index and Palliative Prognostic Score were considered documentation of a PC-specific EBP.13–15 Furthermore, based on noncancer comorbidities, specific EBP was also evaluated in the electronic medical record. Key signs, symptoms, or biological parameters (e.g., anorexia, dyspnea, hypercalcemia, etc.) associated with survival were not evaluated, as determination of these parameters as part of routine clinical assessment versus prognostication could not be determined retrospectively. All consultation notes included in this analysis were reviewed by two of the coauthors.
Statistical analysis
Summary statistics are reported for patient demographic data and for prognostic indicators. Differences in age and ethnicity versus the reason for consultation were evaluated by Kruskal-Wallis, chi-squared, and Fisher's exact test (for n≤5), where appropriate. All statistical analyses were performed with statistical software SAS (SAS version 8.0; SAS Institute Inc., Cary, NC).
Results
The sample size was 412 oncology patients, with a distribution of 176 women and 236 men. Mean age±SD was 58±16 years. Most patients were Caucasian, with gastrointestinal cancer the most common (see Table 1). There was no difference in ethnicity between cohorts. Documentation of functional status, PC-specific EBP, and other prognostication models/calculators are summarized in Table 2. Out of 412 consultation notes, 24 (6%) had documentation of functional status, with ECOG being the most common (n=20). There was no documentation of the Palliative Performance Scale in all of the reviewed consultation notes. Out of 412, 4 (1%) consultation notes contained EBP. The Model for End Stage Liver Disease (MELD) Score was documented for three patients with hepatocellular carcinoma, and the Walter Index and the Lee Index were used for an elderly patient with prostate cancer. In the absence of a PC consultation note template, documentation of the Palliative Prognostic Score and Palliative Prognostic Index was absent.
The Walter Index and the Lee Index were documented for the same patient.
ADLs, Activities of Daily Living; ECOG, Eastern Cooperative Oncology Group; KPS, Karnofsky Performance Scale; MELD, Model for End Stage Liver Disease.
Discussion
Formulating, discussing, and documenting prognosis should be a core skill for PC clinicians. 16 However, prognostication is highly varied among providers and institutions. 17 Prior studies demonstrate EBP yields more accurate prognostication, yet broad implementation of standard EBP practices is lacking.6,8,18 Despite specific consultation to address goals of care, in half of the consultations, documentation of EBP was 1%. This retrospective study demonstrates PC providers need to improve documentation of EBP at our institution. In contrast, Zibelman and colleagues observed a 70% survival estimate documentation rate in the PC note. 3 There may be several reasons for this discrepancy between studies. First, Zibelman and colleagues 3 utilized a PC consultation note template, prompting PC providers to document prognosis. Secondly, documentation of prognosis in broad time intervals such as “hours to days” or “days to weeks” may be more familiar for PC providers than EBP. Calculating and documenting a functional assessment or EBP requires knowledge of and familiarity with these scales, which when lacking may limit their use and documentation. For example, PC providers may not be aware of differences between the Palliative Prognostic Index (median survival in terminally ill cancer patients) and Palliative Prognostic Score (median survival in terminally ill patients) and in which settings each is validated and should be implemented.13–15
We acknowledge several limitations to this study. The primary limitation is the retrospective design limited to a single institution. Consequently, broad conclusions are restricted and causative associations with lack of EBP documentation cannot be determined. We also recognize that other PC programs routinely include EBP as part of their standard initial PC consultation note. However, prognostic documentation appears to be highly varied between institutions. We suspect our experience is not isolated and chose to highlight it to promote discussion and to seek out potential solutions.
In the age of the electronic medical record there is an opportunity to improve our prognostic documentation. In pursuing further prospective studies, a two-pronged approach designed to address the lack of familiarity and inconvenience of documentation of EBP should be addressed. We plan to implement a standard template for the PC consultation note requiring prognostic documentation and to integrate functional assessments and EBP within the electronic medical record. We are working on a prospective study evaluating the use of a website containing over 60 disease-specific EBP reports. We hypothesize that providing a single clinically meaningful and convenient prognostication resource will improve both familiarity with and documentation of EBP. Ideally, increasing familiarity with and use of EBP will in turn improve communication surrounding this sensitive and important topic.
Footnotes
Author Disclosure Statement
No competing financial interests exist.