The association between emergency medical services staffing patterns and out-of-hospital cardiac arrest survival
Posted Apr 12 2010 11:21pm
A recent study looks at The association between emergency medical services staffing patterns (the number of paramedics dispatched on an ambulance) and out-of-hospital cardiac arrest survival. They make some interesting observations.
We tested the hypothesis that two or more paramedics at the scene of OHCA would be correlated with a higher rate of survival to hospital discharge. 
Since there were at least 2 paramedics on scene, they really compared the presence of 2 medics with the presence of 3 medics and with the presence of more than 3 medics. OHCA is Out-of-Hospital Cardiac Arrest.
Paramedic treatment of cardiac arrest is provided by protocol and direct medical oversight.
An interesting interpretation of the word direct. A medical command physician is available by phone or radio for consultation/orders. This is historically the way direct medical oversight has been interpreted, but it requires some flexibility in the interpretation of the meaning of direct. The word oversight requires similar flexibility, since the physician depends entirely on what the medic says, except for the medical command physicians who perversely just say no to everything.
Cardiac arrests resulting from a drug overdose, suicide, drowning, hypoxia, exsanguination, stroke, or trauma were excluded from the study. Also excluded were cases in which no crew configuration or responding unit information was available, cases in which no resuscitation effort was attempted, and cases in which no time data were available.
It is reasonable to exclude some of these causes, since they are not standard medical cardiac arrests and do not contribute significantly to resuscitation statistics. The intent of the study seems to be the effect of the number of medics on standard cardiac arrest. Trauma is reasonable to exclude, since resuscitation from traumatic arrest is, for all practical purposes, zero percent. Movies and TV suggest otherwise, but there is nothing about chest compressions that reverses blood loss, TBI (Traumatic Brain Injury), pneumothorax, cardiac tamponade, or other traumatic causes of sudden death. On the other hand, there are some causes of cardiac arrest that do respond to chest compressions, such as commotio cordis  ,  or lightning strike,  that might appear to fall into the trauma category. Cardiac arrest due to stroke may have resuscitation outcomes as dismal as trauma.
Why exclude cardiac arrest due to hypoxia? Are they looking only at airway obstruction as a cause of hypoxia? Does this include anaphylaxis or asthma? What about CHF (Congestive Heart Failure), which could be categorized as hypoxic or as cardiac? These are questions that were not addressed.
There is no explanation for any of these exclusions anywhere in the paper. Are these excluded because the authors consider resuscitation to be so unlikely that they would only contribute a more zeros to the data?
If a paramedic responded as part of the ﬁrst response unit’s crew, the paramedic was not counted as being part of the crew conﬁguration. The independent variable was categorized solely on the number of paramedics present in the responding ALS unit crew. Milwaukee County EMS operates with a minimum of two paramedics on ALS ambulances. During the study period, there were no cases treated by one paramedic.
In other words, although we are looking at whether more medics lead to better outcomes, we will set aside cases where the medics are first responders, unless those first responders are responding on an ambulance. Interesting, but it is difficult to tell if this would have any influence, since no numbers are reported. If medic first responders are rare, it probably will not have any effect on outcomes. Not having data, we do not know if it is rare. If medic first responders are common, this should not be ignored as a variable.
It might be nice to evaluate the effect of paramedic first responders. Even though there does not appear to be any benefit from having paramedics treating cardiac arrests, as first responders or otherwise, some systems are adopting the practice of having every responder be a paramedic. These systems seem to encourage the belief that, even though paramedic treatments do not improve outcomes, the sight a lot of paramedics on scene is more important than anything that might actually improve outcomes. There is no need to address things that matter, when we can put on an impressive show, unless survival is important.
Neurologic status at discharge was not available.
Too bad. That could be useful information to have.
There were sufficient data to analyze 10,057 (98%)cases.
That is very good.
FIGURE 2. Frequencies of cardiac arrest outcome by year in Milwaukee County. ROSC = Return Of Spontaneous Circulation. I added the red circles and green squares to make it easier to identify the lines. Nothing else was changed about this chart.
The most interesting thing that I notice is that increases in one of the surrogate end points (ROSC) seems to indicate drops in the end point that matters - survival to discharge. While they did not have access to the neurological status at discharge, this information seems to contradict what everyone claims is important about resuscitation - If we don't have ROSC, we cannot improve survival to discharge.
In the above chart, as ROSC increases, survival to discharge decreases. Is this statistically significant? I don't know, but it appears to be pretty consistent. The numbers are not provided by years, but the trend can be determined from the chart. Below, I list the the changes from year to year in ROSC (Return Of Spontaneous Circulation) and in Survival (survival to discharge from the hospital). ⇓ = year to year decrease. ⇑ = increase. ⇔ = no change. More = more.
1993 to 1994 ROSC ⇓ Survival ⇓
1994 to 1995 ROSC ⇓ Survival ⇑≠
1995 to 1996 ROSC ⇑⇑⇑⇑ Survival ⇓≠
1996 to 1997 ROSC ⇑⇑ Survival ⇓≠
1997 to 1998 ROSC ⇑⇑ Survival ⇔≠
1998 to 1999 ROSC ⇓⇓⇓ Survival ⇓
1999 to 2000 ROSC ⇑⇑ Survival ⇑
2000 to 2001 ROSC ⇓⇓⇓ Survival ⇑≠
2001 to 2002 ROSC ⇑⇑ Survival ⇓⇓≠
2002 to 2003 ROSC ⇑⇑ Survival ⇑⇑
2003 to 2004 ROSC ⇓ Survival ⇓
2004 to 2005 ROSC ⇓ Survival ⇑⇑≠
I need to point out that this study was not designed to examine any connection between ROSC and survival to discharge. The yearly data are not included, so I am only looking at the direction of change of the bars connecting one year to the next. Out of 12 years, the change in percentage of ROSC is the same as the change in percentage of survival only 5 times. Sometimes these divergences are dramatic. Almost every big change in ROSC had an opposite change in survival.
As percentage of ROSC improves, percentage of survival seems to decrease. Maybe we need to stop obsessing about improving ROSC and just work on the more complicated problem of improving long term survival, which is all that really matters.
it appears that even though there was a medication change in the treatment protocol, changes to the American Heart Association guidelines, advances in abilities, training, equipment, CPR performance, and variation in hospital care, survival to hospital discharge remained stable during the study period. This may indicate that we have not yet identified the factors that are crucial to improving survival and that more research is needed to find the ideal treatment for cardiac arrest.
Contrariwise, it may indicate that we have already found the most effective paramedic/ALS (Advanced Life Support) treatment. We are just unwilling to accept it, because we cannot believe it is that simple. Excellent continuous compression CPR interrupted only by rapid defibrillation.
We do not need paramedics for this. Therefore the number of paramedics on scene may only lead to interference with effective treatment.
It is important to note that, as is shown in Table 1, crews with two paramedics treated fewer cardiac arrest cases with an initial rhythm of ventricular fibrillation or pulseless ventricular tachycardia than crews with three or more paramedics. Yet the unadjusted and adjusted odds ratios demonstrated that two paramedics conferred a survival advantage. This seems counterintuitive and may indicate an even stronger association between crew size and survival.
If there is a bias in the data, it is likely one that hides the magnitude of the harm to patients from more paramedics.
The Milwaukee County EMS system operates with a minimum of two paramedics on ALS ambulances. During the study period, no cases were treated by one paramedic. A single paramedic’s inﬂuence on outcome was not able to be evaluated.
The data do appear to be stating that the more paramedics on scene, the less likelihood that the cardiac arrest patient will leave the hospital alive.
When adjusted for variables previously correlated with cardiac arrest survival and referenced against crews with two paramedics, patients treated by crews with three paramedics (0.83, 95% conﬁdence interval [CI] 0.70 to 0.97, p = 0.02) and crews with four or more paramedics (0.66, 95% CI0.52 to 0.83, p < 0.01) were associated with reduced survival to hospital discharge.
The number of paramedics does not appear to interfere with ROSC, only with long term meaningful outcome.
The studies of ALS interventions keep pointing out the increased ROSC with ALS interventions. Unfortunately, the surrogate endpoint of ROSC does not appear to lead to improved survival to discharge.
Standard cardiac arrest treatment has led to more of an emphasis on continuous compressions and defibrillation. Standard cardiac arrest treatment still includes epinephrine, amiodarone, atropine, intubation, and intravenous access - treatments that have repeatedly failed to show improved survival to discharge.
During cardiac arrest, basic CPR and early defibrillation are of primary importance, and drug administration is of secondary importance. Few drugs used in the treatment of cardiac arrest are supported by strong evidence. After beginning CPR and attempting defibrillation, rescuers can establish intravenous (IV) access, consider drug therapy, and insert an advanced airway. 
By supported by strong evidence, they mean that there is not any evidence of improved survival to discharge with drugs, IVs, endotracheal tubes - anything other than good continuous compressions and defibrillation.
Maybe this study means that the more people capable of performing less-than-helpful treatments, the more likely that less-than-helpful treatments will be given.
Perhaps, with the next revision of the cardiac arrest guidelines, we will make the ethical decision to limit treatments to those that have been shown to improve survival, rather than subjecting everyone to these treatments that have not been shown to improve survival.
^1The association between emergency medical services staffing patterns and out-of-hospital cardiac arrest survival. Eschmann NM, Pirrallo RG, Aufderheide TP, Lerner EB. Prehosp Emerg Care. 2010 Jan-Mar;14(1):71-7. PMID: 19947870 [PubMed - indexed for MEDLINE]
^2An experimental model of sudden death due to low-energy chest-wall impact (commotio cordis) Link MS, Wang PJ, Pandian NG, Bharati S, Udelson JE, Lee MY, Vecchiotti MA, VanderBrink BA, Mirra G, Maron BJ, Estes NA 3rd. N Engl J Med. 1998 Jun 18;338(25):1805-11. PMID: 9632447 [PubMed - indexed for MEDLINE]
^3Protecting our children from the consequences of chest blows on the playing field: a time for science over marketing. Link MS, Bir C, Dau N, Madias C, Estes NA 3rd, Maron BJ. Pediatrics. 2008 Aug;122(2):437-9. No abstract available. PMID: 18676560 [PubMed - indexed for MEDLINE]
^4Electric Shock and Lightning Strikes 2005 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care Part 10.9: Electric Shock and Lightning Strikes Free Full TextFree PDF
^5Management of Cardiac Arrest 2005 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care Part 7.2: Management of Cardiac Arrest Access for Medications: Correct Priorities Free Full TextFree PDF