Qualitative Factors in Patients Who Die Shortly After Emergency Department Discharge

Abstract

Early death after emergency department (ED) discharge may signal opportunities to improve care. Prior studies are limited by incomplete mortality ascertainment and lack of clinically important information in administrative data. The goal in this hypothesis-generating study was to identify patient and process of care themes that may provide possible explanations for early postdischarge mortality. This was a qualitative analysis of medical records of adult patients who visited the ED of any of six hospitals in an integrated health system (Kaiser Permanente Southern California [KPSC]) and died within 7 days of discharge in 2007 and 2008. Nonmembers, visits to non–health plan hospitals, patients receiving or referred to hospice care, and patients with do not attempt resuscitation or do not intubate orders (DNAR/DNI) were excluded. Under the guidance of two qualitative research scientists, a team of three emergency physicians used grounded theory techniques to identify patient clinical presentations and processes of care that serve as potential explanations for poor outcome after discharge. The source population consisted of a total of 290,092 members with 446,120 discharges from six KPSC EDs in 2007 and 2008. A total of 203 deaths occurred within 7 days of ED discharge (0.05%). Sixty-one randomly chosen cases were reviewed. Patient-level themes that emerged included an unexplained persistent acute change in mental status, recent fall, abnormal vital signs, ill-appearing presentation, malfunctioning indwelling device, and presenting symptoms remaining at discharge. Process-of-care factors included a discrepancy in history of present illness, incomplete physical examination, and change of discharge plan by a third party, such as a consulting or admitting physician. In this hypothesis-generating study, qualitative research techniques were used to identify clinical and process-of-care factors in patients who died within days after discharge from an ED. These potential predictors will be formally tested in a future quantitative study. La muerte precoz tras el alta del servicio de urgencias (SU) puede indicar oportunidades de mejora en la atención. Los estudios previos están limitados por la determinación incompleta de la mortalidad y la ausencia de información clínica importante en los datos administrativos. El objetivo de este estudio generador de hipótesis fue identificar aspectos del paciente y del proceso de atención que pueden proporcionar posibles explicaciones para la mortalidad precoz tras el alta. Estudio cualitativo de las historias médicas de los pacientes adultos que visitaron cualquiera de los SU de los seis hospitales integrados en un sistema sanitario (Kaiser Permanente Southern California [KPSC]) y que fallecieron en los primeros 7 días tras el alta en 2007 y 2008. Se excluyeron los no miembros, las visitas a los hospitales no pertenecientes al sistema sanitario, los pacientes recibidos de, o referidos a, cuidados paliativos y los pacientes con órdenes de no intubación o no resucitación. Bajo la guía de dos científicos de investigación cualitativa, un equipo de tres urgenciólogos usó las técnicas de la teoría fundamentada para identificar las presentaciones clínicas de los pacientes y los procesos de atención que sirven como explicaciones potenciales para los malos resultados tras el alta. La población del estudio consistió en un total de 290.092 miembros con 446.120 altas desde los seis SU del KPSC en 2007 y 2008. Un total de 203 muertes ocurrieron en los primeros siete días tras el alta (0,05%). Se revisaron 61 casos elegidos de forma aleatoria. Los aspectos relativos al paciente que emergieron incluyeron un cambio agudo en el estado mental persistente e inexplicado, una caída reciente, una alteración de las constantes vitales, una apariencia de enfermo a la llegada, un mal funcionamiento del dispositivo y una presencia de síntomas que permanecen en el momento del alta. Los factores del proceso de atención incluyeron una discrepancia en la historia de la enfermedad actual, una exploración física incompleta y un cambio del plan al alta por un tercero, como un médico de admisión o un consultor. En este estudio de generación de hipótesis, se usaron técnicas de investigación cualitativa para identificar los factores clínicos y del proceso de atención en los pacientes que fallecieron en los primeros 7 días tras el alta de un SU. Estos factores predictivos potenciales serán formalmente comprobados en un futuro estudio cuantitativo. Early mortality after emergency department (ED) discharge may represent opportunities to improve diagnosis, treatment, and postdischarge care. Contributing factors to such events may include the inherent high-risk nature of ED care,1 ED crowding,2-9 and an increasing illness severity of the U.S. population.3, 4, 10 Current estimates suggest that 41,500 patients discharged from EDs die within 7 days annually in the United States.11, 12 Although early death after discharge from the ED has important patient safety and quality improvement implications, there is insufficient understanding of such events. Prior U.S. studies are limited by an incomplete capture of deaths.13, 14 Our study team recently analyzed predictors of early death using administrative data; however, that study omitted critically important information on clinical and process factors that can only be abstracted through chart review.11 The objective of this study was to identify patient and process-of-care factors that may be associated with early death after discharge. We performed a qualitative chart review to identify potential predictors that are not measured by existing administrative data. This methodology is well suited for generating hypotheses that can be formally tested in future studies. This was a qualitative chart review analysis of patients who died within 7 days of ED evaluation and discharge. The study protocol was approved by the institutional review boards of Kaiser Permanente Southern California (KPSC) and the University of California at Los Angeles. Study subjects were members of KPSC, an integrated health system that provides comprehensive care to 3.5 million members throughout Southern California. Health care is delivered at over 100 outpatient clinics, and emergency services are provided at 14 medical centers. Electronic ED charts were available from six health plan medical centers; the remaining eight EDs had not completed implementing the electronic record system during the study period. All health plan members have similar health care benefits, including coverage of emergency services both within and outside the health system. Electronic administrative databases track all health care encounters within the health system and provide access to notes, vital signs, laboratory, pharmacy, imaging, and many other clinical care data. A claims reimbursement system tracks health care encounters at outside facilities. A specialized administrative database provides information on the use of hospice services. Study subjects were patients who died within 7 days of discharge from a health plan ED between January 1, 2007, and December 31, 2008. To arrive at our cohort, administrative data were used to identify member visits to any of the six EDs, and death within 7 days of the ED visit and discharge was identified using the California Vital Statistics files and the Social Security Death Index. We chose the 7-day time frame because of its clinical relevance, implications for health policy decisions, and prior use in related studies.11, 13-16 Each included subject had to be a member of the health plan at the time of the ED visit; however, no minimum health plan enrollment period was required. Subjects were restricted to age 18 years and older because of the inherent difference between pediatric and adult presentations and outcomes. Patients seen and discharged from the ED to home or a non–acute care facility (nursing or rehabilitation) were eligible. Discharges to non–acute care facilities were considered eligible as these facilities lack either on-site or off-site continuous physician coverage and diagnostic abilities and require patients be transferred to EDs for acute evaluations. Patients admitted to inpatient or observation status beds were excluded. Patients in hospice care were excluded on the assumption that this status implies an intention to provide end-of-life comfort therapy rather than to prolong life. Members who visited non–health plan EDs were also excluded, as we did not have ED charts for these patients. From the 203 patients who died within seven days after ED discharge, we randomly assigned a rank order and proceeded down the list with a goal of reviewing a set of charts to arrive at preliminary themes through theme saturation (Phase 1) and a set of charts to validate the previously identified themes (Phase II). A total of 64 charts were extracted, but three were excluded during manual chart review because they were found to have do not attempt resuscitation/do not intubate (DNAR/DNI) orders in the chart or were offered hospice services during the ED visit. The final analysis involved the review of 61 charts (30 in Phase I and 31 in Phase II). Deidentified printed medical records were reviewed by three emergency physicians (EPs; GZG, NJM, DF) using grounded theory techniques. Grounded theory is a form of qualitative analysis that uses social research methods to acquire theories from data. Chart data included the emergency physician (EP), nursing, and consultant records; the reported laboratory and imaging findings; and the most recent primary medical doctor visit prior to each ED visit. The most recent primary medical doctor visit gave reviewers insight into each patient's condition prior to the ED visit in question. Using the qualitative methods and coding techniques described below17-19 and after receiving training in qualitative methodology, the three EPs reviewed charts both independently and as a team to identify a set of coded themes. Training of the reviewers involved multiple meetings to discuss the theory and practice of qualitative methodology14, 18-20 and to sample chart reviews with the team lead qualitative scientist (ST). Reviewers were advised to assess each chart based on the written information present; for example, if there was no documentation or mention of the vital signs, the chart was assumed to lack vital signs. We used grounded theory to identify salient themes such as patient presentations and processes of care to build a conceptual framework that provides possible explanations for death within 7 days of ED discharge. The process of data collection and analysis was based on prior literature18, 19 and the recommendations of our team qualitative researchers (ST, GWR). The group met monthly to discuss and refine the codes, often described as single words such as “fall,” into preliminary themes, defined as broader categories such as “fall in elderly.” This process continued until thematic saturation was achieved at approximately 30 charts, and no new concepts were being generated. To ensure that we had achieved thematic saturation, a second set of charts was reviewed (n = 31) using the same iterative grounded theory methodology. Throughout the reviews, the team lead qualitative researcher (ST) regularly evaluated the coding scheme to ensure that the observations were comprehensive and that narrative accounts were being generated. Per qualitative research guidelines,18, 19 the final data were then reviewed by the study team (GZG, SFD, BCS) to ensure that the interpretations were credible. With qualitative expertise from fields in sociology (ST) and anthropology (GWR), the preliminary themes were sorted by the study lead author (GZG) into a final set of analytic themes that could potentially explain early death after discharge. All cases had multiple themes. For each case, we assigned a single theme that was felt to be the most important contributor to death after ED discharge. The process of categorizing the themes involved twelve inductive data reviews (defined as reviews of the charts that derived overarching analytic themes from preliminary themes) and discussions of the analytic themes with the study team to arrive at final themes that were thought to be the most probable contributors to death after ED discharge. To qualify as a final theme, at least three cases needed to have the same finding. The analysis yielded 11 final themes that were categorized as occurring either at the patient level or the process level. Coding data were managed with Microsoft Excel (Microsoft Corp., Redmond, WA) Descriptive statistics for the cohort were obtained using SAS 9.2 (SAS Institute, Cary, NC). Over the 2 years, there were 497,996 patients who accounted for 813,232 visits to six EDs (Figure 1). After exclusions, the source population contained 290,092 health plan members with 446,120 ED visits. The characteristics of the source population and their comorbidities are presented in Table 1. The mean (±SD) age of the source population was 46.8 (±18.2) years. A total of 203 patients died within 7 days after ED discharge, with a mean (±SD) age of 71.1 (±14.7) years. Of the patients who died, 61 were included in the final analysis (Figure 1), of which the mean (±SD) age was 72.0 (±12.5) years. The themes were classified into patient- or process-level categories. Patient factors describe characteristics of patients, their medical history, or their presentation in the ED. Process-of-care factors are defined as characteristics of the care the patients received during their ED stays. Each case that had overlapping themes was categorized in the one unique theme that the reviewers believed was most likely associated with death after ED discharge. Table 2 summarizes the themes and provides example cases. Of the 61 charts reviewed, four did not have a theme that fell into a category. These cases most likely had deaths that were not related to the ED presentation, such as being hit by a car or suffering an accidental death. The mental status change was identified by either someone accompanying the patient or the EP and could not be explained based on the acquired history, examination, or results of ancillary tests. In addition, all patients with this presentation continued to be confused at the time of discharge. A history of a recent fall in a patient age 65 years or older, occurring within the week prior to the ED visit, in most cases was attributed to mechanical causes without mention of associated symptoms. Abnormal vital signs were defined based on clinical consensus and previous literature15 as heart rate greater than 99 or less than 60 beats/minute, systolic blood pressure greater than 179 or less than 90 mm Hg or diastolic blood pressure greater than 109 mm Hg; respiratory rate greater than 20 or less than 12 breaths/minute; temperature greater than 37.9°C; and oxygen saturation less than 93%. Cases in this category often had initial abnormal vital signs that persisted throughout the ED visit despite interventions. Also, reasons for the abnormal vital signs were seldom addressed in the ED note. The vital sign abnormality most common to the cases was tachycardia. The determination of ill appearance was based on documentation of or in the of the physical All of these patients also were found to have abnormal of either respiratory or indwelling was defined as an not contained in the such as a or The of the presenting was based on documentation of a by the or the physician at the time of discharge. In most the provided a comprehensive of the patient's condition throughout the ED and at discharge. The discrepancy occurred as a of or information to the EP note. In all these the patient prior to medical or during the in the ED and there is no in the ED of the third incomplete was defined as an that did not the and of the that was to the For example, if the patient of an incomplete include the of a Cases with potential were often by laboratory or imaging results that the for a Cases that were found to have their level had between their clinical and objective such as vital signs, laboratory, or imaging The change either by an inpatient admitting physician of or by the patient to medical of In the Kaiser the for an can be by the but the final to is by the admitting physician. In all the EP in or that the patients We identified a set of and process-level factors that ED encounters that in early death after discharge. Our hypothesis-generating study not only provides and quality improvement with important information high-risk but the for future studies. The goal of this research is to such as a patient discharge or electronic at discharge, to early deaths after discharge. case studies and one administrative data study death after ED discharge have in years. Using data and qualitative found cases with the abnormal vital signs, presentation, of a and abnormal mental A case of charts by using medical data found related deaths to be by or heart and studies were limited by incomplete deaths that were not reviewed by the or medical have excluded. We previously reported an analysis of administrative data using the same source cohort as the We found that 7-day death after ED discharge of the time of The predictors were associated with death after older and the three primary discharge of and heart The study on our prior by potentially important patient and process factors that are not by administrative data. In this study, we found patient presentations to be present in cases of early mortality after discharge. A common theme was an unexplained persistent acute mental status change in any age adult or recent fall in an mental is often to illness, or system or This is with prior The of this theme that EPs may the level of confused patients, as a of a lack of to assess or a lack of ancillary tests. are a of in U.S. studies have found that an of mortality in the often fall to or poor The identified in our study were often attributed to mechanical causes with information acquired in the note. Our suggest that an ED a understanding of the patient's physical to the fall, with a of the fall We found abnormal vital signs to be a theme in our Although this a potentially high-risk Of all vital signs, we identified to be the most common abnormal vital In all the abnormality in vital signs either persisted despite or there was no of This the of vital signs, the of a diagnosis, and the to the of the We also found the malfunctioning of an indwelling to potentially an poor We that this is a common that often not illness, but our suggest that the evaluation of such a patient be theme we identified was the of symptoms at discharge. A persistent such as could or treatment, or care the ED visit, such as or We identified processes in the of diagnosis, and A history of present illness and a discrepancy in between was thought to to potential In most this discrepancy was between the EP and ED In the KPSC electronic health record by each of the Our suggest that ED with all health care involved in patient care and a review of all documentation prior to patient discharge. A was a the EP identify one potential of the and to the the initial was not were cases the to the level of illness, despite objective Patients often present with a of physical and These considered were then For a of patients, the plan of was by a third party, including an admitting physician or a consultant or by the patient medical In the Kaiser Permanente health all patients are by consulting admitting physicians prior to The admitting physicians have the to discharge patients if they Our suggest that over patient may a possible for death. This qualitative analysis a group and was not to but to hypotheses for future studies. We the of as reviewers were that all of the study patients had We also that despite hospice or patients, there may have patients who had In addition, the analyzed cohort was similar to the patients who were not analyzed on measured the analyzed cohort had a rate of This may have potentially our to identify process factors unique to patients with We are a study, with to to formally the between the themes identified in this with poor after ED discharge. our team did not the potential between a theme and the postdischarge of death. We did not physicians involved in the cases to we did not the and their thought processes the We assumed that chart documentation that a process was not performed or that the evaluation did not the we that chart documentation may not capture all events and processes that during a patient We performed our study in an integrated health system and our results may not to other patients, who may potentially be at greater for death discharge from the are in our study. In addition, health plan members may have access to outpatient that may not be available to the and as a the practice of EPs in our study EDs may from of other health Our be in other We that previous in this reported on the patient of death and the death was related to the ED We did not of death information to the of of death and the of information in our study In addition, we felt that an of the of ED care with death be we did not attempt to a of death or a of medical we on the factors common to patients and processes of care in patients who died after ED discharge. We also that in to the 7-day time we could have evaluated in either or time We a qualitative analysis to identify patient characteristics and processes of care that may be seen in patients who early death after ED discharge. Our hypothesis-generating results provide insight for and quality improvement these patients. These the for future to the themes identified in this and to at death.