Weekly Review #4

There was a Critical Care article by Kristensen et al., addressing the correlation between SBP and mortality. This was a retrospective cohort study of patients from a Danish ER and ambulance service between 1995-2013. The primary outcome was to determine a SBP threshold that could predict 7 day mortality, secondary outcome was 30 day mortality. A total of 112 727 patients greater than 18 years of age with a blood pressure measurement were included. They determined that SBP <92mmHg was associated with a 10% chance of 7 day mortality in ED patients, for ambulance patients 7 day mortality rate of 10% was predicted at 87mmHg. These threshold values are congruent with other studies that have examined the association between SBP and mortality. This study however noted that mortality rates began to increase at a SBP of 110mmHg, and that by substituting 90 with 110mmHg, as a hypotension cut-off, would detect one additional mortality per 17 false positives, which is a fair trade off for the increased rate of false positives. Although this is statistically significant it may not be that clinically significant as the positive predictive value, of 7 day mortality, at this threshold is still only 8%. Furthermore the applciation of these cutoffs to a specific patient is hampered by the lack of subgroup analysis: there was no separation of trauma from medical patients, no discussion of injury or illness severity, age, or comorbidities and the effect these may have on pre-test probability of death. There is no discussion of why some patients had blood pressure assessments while others did not, although it can be assumed that those who did not were likely younger and healthier. Because of these limitations it is not only hard to apply these SBP cutoffs to the ED population as a whole; but to cohorts within the general population. From a nursing perspective this research would suggest that the statistical cutoff for increased mortality may lie closer to SBP of 110mmHg than 90mmHg. While this may be of limited clinical significance it does add to a growing body of evidence that challenges the traditional hypotension cutoff values. More importantly however it also suggests that a SBP that falls within a "normotensive" range should not necessarily be a reassuring finding in the presence of illness or injury.      

http://ccforum.com/content/19/1/158



Steve Carroll on emDocs had a great piece discussing the EM mindset. A celebration of ED medicine as the art of mastering the undifferentiated patient. He discusses strategies for the three stages where a clinician can improve their practice: before, during and after shift. Before Shift: a commitment to lifelong learning and a pre-work process of mentally preparing for work; During Shift: he discusses the clinical approach, recognizing personal biases, person needs, team work, assumptions about patients wishes, how to advocate for patients and how to prepare for poor outcomes. Aftershift: the need for activities outside of professional interests, and the importance of physical activity for mental health. This is a great piece for anyone working in an emergency department. From a nursing perspective I think the message here is directly relevant, take care to be prepared, take care of yourself, and you will be able to take care of others.  

http://www.emdocs.net/em-mindset-steve-carroll-masters-of-the-undifferentiated-patient/



Jones et al., published a review of unfinished nursing care in the International Journal of Nursing Studies. Unfinished nursing care is an example of health care "under-use", is a form of medical error, and may account for more errors than overuse and misuse of healthcare combined. This review covers original research, psychometric analyses, and review articles on the subject from CINAHL and MEDLINE. 1828 articles were identified, 47 met inclusion criteria, 89% of which defined unfinished nursing care as: tasks undone, implicit rationing, or missed care. There was no substantive difference in definitions, or organizational structure between studies, and all of the research discussed both: nurse work environment, and patient safety/outcomes. 14 survey method articles assessed an inventory of nursing care activities left undone and the reasons why, the majority used a 5 point likert style instrument, on a convenience sample of nurses. Results indicate that on average nurses are only rarely or occasionally (2-3 on a 5 point likert scale) leaving care unfinished, and that it has a high prevalence, with 55-98% of nurses leaving one or more tasks undone. The top 5 unfinished tasks are: emotional support, education, care coordination and discharge planning, care planning, and timeliness of care. Task least frequently missed were: infection control, tests, treatment, procedures, nutrition and elimination. Tasks that addressed psychological as opposed to physiological needs, that had less immediate effect, that required more time and that were less likely to be audited where more likely to be omitted. Although staff ratios, patient acuity and labour resources were poorly linked with increased levels of unfinished care; they were the most commonly cited reason for unfinished care. There was little direct measurement of the effect of missed care on patient or nurse outcomes; but most nurses identified that missed care contributed to worse patient outcomes and decreased levels of nurse satisfaction. There was only one study that examined interventions to decrease unfinished work, the study intervention consisted of podcasts, education, and staff role playing. The intervention had a statistically significant impact on staff perception of teamwork; but failed to result in a clinically significant improvement in teamwork (4-6%). This research establishes frameworks for assessing unfinished care and discusses the impact of missed care. It identifies that current research is weak, that the evidence is mostly self reported, under-powered, and from only three researchers. It is therefore hard to draw causal inferences between the barriers and work left unfinished. From a nursing perspective it's heartening to see quantitative research being performed on daily tasks, it certainly draws attention to work being left undone, and would warrant some reflection by nurses: what tasks are being left undone, is this fair prioritization of effort, and what can be done to address the most pressing issues first?

http://www.sciencedirect.com/science/article/pii/S0020748915000589
.


Venkatesh et al., performed a prospective observational study of handoffs between physicians in an emergency department. Data was collected by direct observation, and chart review. Primary outcome was inclusion or omission of abnormal vital signs (hypoxia/hypotension). Secondary measures were clinician training level, time of day, length of rounds, ED occupancy rate, and number of interruptions. There were a total of 1163 patient handoffs observed, during handoff reporting of hypoxia (SpO2 < 92%) was omitted 74% of the time; and hypotension (SBP <90mmHg) 56% of the time. Providers were interrupted in 49% of hand-offs. There was no direct association between occupancy rate (a corollary for ED crowding), length of rounds, time of day and omission rates. Senior residents had higher omission rates than junior residents or attending physicians. While the absolute values for this data are interesting, and would seem to be congruent with anecdotal rasons for why omissions occur this study was underpowered and none of the secondary measures reached statistical significance: the odds ratio of secondary measures resulting in hand-off errors all crossed 1. Unfortunately this study is of limited statistical use as it is under-powered, it doesn't discuss if all cases of hypoxia or hypotension were clinically relevant, part of expected clinical course (and thus not needing mention), or if these omission did, or could, lead to adverse outcomes. Of interest however it does note that crowding appears to have limited impact on omission rates, and that providers are interrupted (Primarily by nurses) in a large number of cases. From a nursing perspective this would seem to suggest that there is merit using a standardized approach for handing over patients for all members of the team; and that perhaps we should be mindful of interruptions, something I'm sure all nurses are intimately familiar with.

http://www.sciencedirect.com/science/article/pii/S0196064415001845



There was a post on boringEM by Michael Garfinkle on using likelihood ratios (LR) as metrics for assessing the usefulness of diagnostic procedures. He discusses the background statistics required for determining LR's: Sensitivity, Specificity and pre-test probability. There are case examples used to illustrate how pre-test probability can change the likelihood of disease for both positive and negative tests results. There is also a link to a website and app created by Garfinkle for determining pre-test probabilities. The website allows you to search pre-test probabilities by specialty, diagnostic modality, pathology, or by prevalence rate. Each category will link to a list of physical assessment or investigations and the LR's for each.


http://boringem.org/2015/05/07/how-to-use-likelihood-ratios/

http://lrdatabase.com/index.php