| WS6a | Time of call to time of arrival at scene (response time). Proportion of emergency calls with a response time within an agreed standard | Response time is important but the arbitrary standard is not. Response time is important to patients, in terms of panic and anxiety. Response time is also related to outcome: shorter time = better outcomes, all things being equal. Overall, response time was judged important because of its relationship to the patient. However, if we have another measure in the list that is better, this will supersede this one | We would risk adjust to ensure fair comparisons, for example for an ambulance service which covers a large geographical area compared with a small area. Response times are already compared nationally without risk adjustment; therefore, there is no need to risk adjust here. We should use the mean or median response time. Mean of log or geometric mean | Include in workstream 4, no risk adjustment |
| WS6ei | Proportion of eligible calls who arrive at definitive care within agreed timescales, for example a specialist heart attack centre within 150 minutes, a specialist stroke centre within 60 minutes, a major trauma centre within 45 minutes | This is important as time frame is related to a successful outcome. This could include the 8-minute response time for OHCA | A success is arriving in the right place (alive). It is not necessary to risk adjust if this measure is concerned with getting patients to the right place, but if it is for treatment we may want to case-mix adjust for the proportion that could benefit from treatment:People with these conditions should get to an assessment centre to see if they can benefit from treatment Time is measured until the patient handover end point
| Include in workstream 4, no risk adjustment |
| CM1b | Number of calls prioritised correctly to appropriate level of response as a proportion of all 999 calls |
Should this be the number of all calls prioritised correctly or do we want the number of serious calls that were prioritised correctly?
Calls coded green that are left too long will be picked up in the mean response time measure
Just take the serious emergency conditions (16 conditions)
| | Superseded by CM1c – better measure |
| CM1c | Proportion of life-threatening category A calls correctly identified as category A | Change this to calls for serious emergency conditions and use 16 conditions. Clarify what the 16 conditions are |
We need to adjust for case mix. Some measures are difficult to categorise and others easy, for example ruptured aneurysm or OHCA
We will adjust for case mix to level the playing field
We will adjust for age
| Include in workstream 3, for risk adjustment |
| CM2a | Proportion of all cases with a specific condition who are treated in accordance with established protocols and guidelines, for example stroke, heart attack, diabetes mellitus, falls (specify which of these or other conditions you think are important) | This is current CQI for five conditions and is important (asthma, STEMI, stroke, OHCA?) |
The CQIs are not risk adjusted
The question here is whether or not patients with specific conditions were given the agreed best treatment. Therefore, there is no need to risk adjust
Query – should we send a list of conditions that we are not risk adjusting for to ambulance services to ask them if they were performing badly on this measure, what would their excuse be?
| Include in workstream 4, no risk adjustment |
| PO5c | Proportion of patients with a life-threatening condition who are discharged alive from hospital |
Use the 16 serious emergency conditions as the life-threatening conditions
This is important to patients
What about hospital effects?
HSMR looks at the proportion of patients who die in hospital. Do we need to split this into prehospital and hospital components?
If we have the HSMR we also need the prehospital SMR for all serious conditions. This is the proportion of deaths that occur before admission to a bed
The HSMR is the proportion that die in hospital
The system HSMR is the proportion that die before discharge
| HSMR is case-mix adjusted, so the PHSMR and the SHSMR will case-mix adjusted | Include in workstream 3, for risk adjustment |
| PO1a | Proportion of all patients seen by an ambulance crew who have a pain assessment recorded | This is an explanatory variable for PO1c | | Superseded by PO1c |
| PO1b | Proportion of patients who report pain who are given pain relief | This is an explanatory variable for PO1c | | Superseded by PO1c |
| PO1c | Proportion of patients who have a reduction in pain score after analgesia treatment |
Should this be the mean reduction in pain (adjusted for time)?
Remove after analgesia as other treatments, for example splints, can reduce pain
|
Risk adjust for conditions, age and time
We could start on this with just the ambulance data
There may be high proportions of missing data as this measure requires two pain readings. We need to determine what to do with missing data
We need a model with everything that is outside the ambulance control first
| Include in workstream 3, for risk adjustment |
| PO1d | Proportion of patients reporting pain who have more than one pain score recorded | This is an explanatory variable for PO1c | | Superseded by PO1c |
| WS6e | Time of call to time to definitive care | Too vague. Superseded by WS6ei | | Superseded by WS6ei |
| PO6a | Proportion of all 999 calls re-contacting the ambulance service within 24 hours | The definition of re-contacts needs to be clearer. This should be re-contacts for all patients who were not conveyed to hospital. This is because patients who are conveyed to hospital may be discharged and re-contact the ambulance service | | Superseded by PO6c |
| PO6c | Proportion of patients left at home who have a contact with any emergency/urgent health service within 24 hours |
This supersedes all other re-contact measures. Left at home = ’see and treat’ or ‘hear and treat’
Is a low rate good?
This needs a reciprocal measure about patients who are taken to ED and not treated/admitted
|
Risk adjust for age
Time frame = 24 hours
| Include in workstream 3, for risk adjustment |
| CM2c | Proportion of all cases with a specific condition who meet the established criteria for transfer, who are transported to an appropriate specialist facility, for example a heart attack, stroke or major trauma centre | This is covered by WS6e1 | | Superseded by WS6e1 |
| PO5a_i | Proportion of 999 callers who die within: i. 0–48 hours of first call | We want to look at the proportion of 999 callers who die from specific causes within a specified time frame, where death was avoidable. We need to identify conditions where patients should not die. Look at the cause of death for patients in our sample and identify those that are preventable, for example hypothermia within 24 hours of call | Adjust for age | Include in workstream 3, for risk adjustment |
| WS3b | Proportion of category A calls attended by a paramedic | Not important | | Exclude |
| WS3c | Proportion of patients who are treated on scene or left at home who are referred to an appropriate pathway or primary care |
Is it the non-conveyance rate?
Non-conveyance who are referred
Not important
| | Exclude |
| WS2a | Number of life-threatening (category A) calls not identified as category A as a proportion of all 999 calls | This is important but is the opposite of a measure that is already included (CM1c). Therefore this will be measured as part of CM1c | | Superseded by CM1c |
| WS3f | Proportion of patients who potentially could be left at home who are successfully discharged at the scene | Better measures exist for this. It is difficult to know what the denominator is. We want to look at people who are transported and not having anything done. Superseded by WS3e | | Superseded by WS3e |
| WS3e | Proportion of patients transported to ED by 999 emergency ambulance and discharged without treatment or investigation(s) that needed hospital facilities | This is important and supersedes other measures | ? Not sure | Include |
| CM1a | Proportion of all calls referred for telephone advice returned for a 999 ambulance response | This relates to efficiency. This was highly scored in the Delphi as is easy to measure, but is not important | | Exclude |
| PO3a | Proportion of patients with cardiac arrest where resuscitation is attempted at the incident scene who have a pulse on arrival at the ED |
ROSC is one of the CQIs
There is scope for ROSC to be improved. Use in the 16 emergency conditions as a bundle but need to look at adjusted time to hospital discharge. This will give ROSC, survival to discharge and the system SMR
What about non-transports? Can we pick these up?
| Needs to be adjusted for rhythm, witnessing, bystander CPR and adrenaline | |
| PO6b | Proportion of all 999 calls referred for telephone advice only re-contacting the ambulance service within 24 hours | There are other re-contact measures that are better | | Superseded by PO6c |
| PO6e | Proportion of patients left at home who are admitted to hospital within 72 hours | Left at home = ’hear and treat’ and ‘see and treat’ patients. We are interested in the proportion of patients who are not conveyed who are admitted to hospital. This could be part of a re-contacts set of measures. Include for the moment, but may be dropped | Risk adjust for age, condition type, avoidable emergency conditions. It is impossible to know where the hospital contact is for the same condition, or the extent to which the two are linked; therefore, we will assume that admissions within 3 days are related | Include in workstream 3, for risk adjustment |
| R2_WS6a_2_3_ | | Not important | | Exclude |
| WS6a_1 | Proportion of emergency calls with a response time within an agreed standard for calls for life-threatening conditions | This is dealt with by another measure | | Exclude. Superseded by WS6a |
| WS2b | Number of calls that are not life-threatening identified as category A calls as a proportion of all 999 calls | We have another measure that looks at the opposite side of this (CM1c) | | Exclude. Superseded by CM1c |
| CM1d | Proportion of calls for a specific condition correctly identified at the time of the call, e.g. cardiac arrest, stroke, heart attack | This is about accuracy and recognising conditions. The ambulance AMPDS code is not diagnostic and you would need to relate the AMPDS code to the admission code. This is of low importance and very difficult to verify, and can already be measured using similar measures that are already included. For example, with STEMIs sent to right place | | Exclude. Superseded by WS6Ei and CM1c |
| CM2b | Proportion of cases that comply with end-of-life care plans when these are available | This is important but not currently measurable. ePRFs should record whether or not there is an end-of-life care plan | | Exclude |
| WS6d | Time of call to CPR start time when CPR is required. Average time from call to start of CPR in cases of cardiac arrest | This is explanatory data for ROSC and survival and is a problem with bystander CPR. This is an explanatory process and not an outcome | | Exclude |
| CM1e | Number of people attended within the scope of advanced paramedic practice (treat and leave at home) as a proportion of all people attended on scene | We cannot measure this as we do not know who the paramedics are. This measure is related to efficiency and is not important here. However, this links to policy because if the Keogh report70 becomes standard, more people will be left at home and ambulance services will need to have sufficient paramedics to deal with the demand | | Exclude |
| CM3a | Number of ‘never events’ reported as a proportion of all requests for 999 ambulance care (never events applicable to the ambulance service include administration of drugs by the wrong route and failure to monitor and respond to reduced oxygen saturation) | Never events are already recorded. This does not come into the scope of the PhOEBE programme as not performance measures | | Exclude |
| CM3b | Number of patient safety incidents reported as a proportion of all requests for 999 ambulance care | Same as CM3a | | Exclude |
| PO2c | Proportion of patients who report that key aspects of care were delivered. (Examples of key aspects are timeliness of response, reassurance, professionalism, communication, smooth transition between different services or parts of the same service) | Not measured routinely. No patient satisfaction or experience measures | | Exclude |
| PO2d | Proportion of patients who were satisfied with the overall service and separate components, for example the 999-call handling, attending ambulance crews | Not measured routinely. No patient satisfaction or experience measures | | Exclude |
| PO5d | Proportion of patients with a specific clinical condition (e.g. stroke, heart attack, cardiac arrest) who are discharged alive from hospital | A mass measure would be for the 16 emergency conditions, but we may want to split this for specific conditions. This relates to PO5c and is part of a set | Adjust for case mix? | Include in workstream 3, for risk adjustment |
| R2_WS6a_2-25 | Proportion of emergency calls for conditions that are not life-threatening with a response time of ≤ 25 minutes | This is not a measure | | Exclude |
| R2_WS6a_4 | Proportion of emergency calls for life-threatening conditions with a response time of < 4 minutes (+ 4 minutes exact) | This applies only to cardiac arrest and we are looking at this in another measure | | Exclude |
| R2_WS6a_4_8 | Proportion of emergency calls for life-threatening conditions with a response time of between 4 and 8 minutes | This applies only to cardiac arrest and we are looking at this in another measure | | Exclude |
| R2_WS6E_3 | | No | | Exclude |
| WS1a | Number of completed patient clinical records as a proportion of all cases attended by the ambulance service in accordance with minimum agreed data set | This is a data completeness measure | | Exclude |
| WS3a | Number of calls transferred for telephone clinical advice assessment that are completed with self-care advice or referral to an appropriate service as a proportion of all calls transferred for clinical advice | This is an efficiency measure. This is superseded by other measures | | Exclude |
| WS3d | Proportion of all calls who receive an ambulance response who are not conveyed to hospital or other health service facilities | This is the non-conveyance rate. This is being looked at by VAN | | Exclude |
| WS4a | Proportion of staff who comply with mandatory training requirements for basic and advanced life support (BLS and ALS) | No. This is a service management measure | | Exclude |
| WS4b | Proportion of operational staff trained as paramedics | No. This is a service management measure | | Exclude |
| WS4c | Proportion of paramedics with advanced practitioner training | No. This is a service management measure | | Exclude |
| WS5a | Unit-hour utilisation for the whole service (a unit-hour is a fully staffed ambulance for 1 hour). For any given time period, a service will have multiple unit-hours available. Unit-hour utilisation is how many of those hours are used within that time period |
Unit-hour utilisation works on the premise that calls are spread evenly and there are no spikes
No. This is a service management measure
| | Exclude |
| WS5b | Unit-hour utilisation for urban areas – compared with agreed utilisation | No. This is a service management measure | | Exclude |
| WS5c | Unit-hour utilisation for rural areas – compared with agreed utilisation | No. This is a service management measure | | Exclude |
| WS6a_2 | Proportion of emergency calls with a response time within an agreed standard for calls for non-life-threatening conditions | This will be covered by other measures (WS6a) | | Exclude. Superseded by WS6a |
| WS6e_2 | Proportion of eligible calls who arrive at a specialist stroke centre within 60 minutes | This is covered by CM1c | | Exclude. Superseded by CM1c |
