This study was produced by SCHIN using anonymised data extracted from the Well Close Square patient database. The object of the study was to determine the potential benefit of using PRODIGY 3 at the practice. There are comments by BNC added to the document.

Clinical Audit Feedback

Produced for: Well Close Square Surgery

Produced by: Sowerby Centre for Health Informatics

Circulation: Strictly confidential to Well Close Square surgery

Version: Final

Date: Tuesday, 11 June 2002

Table of Contents

1. Acknowledgements. 3

2. Introduction. 4

2.1. Place of clinical audit in the evaluation of Prodigy Phase 3. 4

2.2. Setting the criteria. 4

2.3. Methodology. 5

3. Results. 6

3.1. Number of patient records for the selected condition. 6

3.2. Asthma audit results. 7

3.3. Angina audit results. 10

4. Conclusion. 16

5. References. 17

1. Acknowledgements

Sowerby Centre for Health Informatics at Newcastle (SCHIN), University of Newcastle, wishes to thank Dr Brad Cheek for the invaluable contribution to the study presented in this report and the provision of data on which the study was based. None of the work could have been done without this.

2. Introduction

This document details the results of retrospective external clinical audit of angina, asthma and hypertension conducted at Well Close Square Surgery (thereinafter “the practice”) by SCHIN, University Of Newcastle for the purposes of Prodigy Phase 3 evaluation.

2.1. Place of clinical audit in the evaluation of Prodigy Phase 3

To overcome numerous difficulties in the evaluation of health informatics interventions and to increase validity, evaluation of Prodigy Phase 3 was designed as a multiple – approach study involving several research techniques. Clinical audit was deemed a very important quantitative approach in this evaluation programme.

Within the initial scope of the evaluation programme it was envisaged that clinical audit would play an important role together with log files and the questionnaire survey in creating a quantitative estimate of the effect of Prodigy Phase 3 intervention. Due to a number of reasons, implementation of Phase 3 was not successful, which essentially invalidates the “change” analysis that was initially planned.

The initial purpose of this clinical audit was to analyse the quality of patient records and prescribing before and after the Prodigy Phase 3 intervention, as well as to collect some information from the electronic patient record essential for Prodigy Phase 3 log file analysis. After the scope of the evaluation programme was revised, the purpose of the clinical audit was changed to providing indicators of potential effect of Prodigy Phase 3. The validity of such an estimate of a potential effect of Prodigy Phase 3 is, of course, limited (ceiling effect, guidance compliance, etc.). It needs to be born in mind that the results only give a general indications of areas and potential sides of improvement. Audit criteria were designed to analyse weather Prodigy Phase 3 could assist this improvement. With the assumption of full compliance with Prodigy Phase 3 guidance it is possible to use the results of the audit to estimate the potential effect of Prodigy Phase 3 use.

Three conditions were chosen as clinical audit targets: angina, asthma and hypertension. This is to coincide with three chronic conditions currently implemented in Prodigy Phase 3.

2.2. Setting the criteria

The knowledge base of Prodigy Phase 3 was a starting point in collating a list of criteria against which the management of patients with the selected chronic conditions was audited. The number of audit criteria was limited, as they had to comply with the following circumstances:

· Resource limitations. The whole process of audit had to be completed in 3.5 months from scratch until the final report by one full-time analyst.

· The criteria had to be “computable”. Due to time limitations it was not possible to consult paper records or clinicians, thus the sources of data were limited to the electronic patient records (EPR). The criteria had to be limited to those computable with a help of SQL queries or various contingency tables with some descriptive statistics.

· Remote analysis. The analysis of the data had to be performed remotely using one-off retrospective data extraction. It was not possible to extract all the data for the necessary patients due to the issues of data protection and storage limitations. Therefore, any analysis that required a full history of medication could not be performed.

The criteria were selected so that the results of the audit would be indicative of the extent to which patient care can be improved if the Prodigy guidance was adhered to. A possible impact of the use of Prodigy Phase 3 on the quality of the recording of the EPR was also analysed.

2.3. Methodology

The following principles were adhered to during the study:

- All the data extracted were used solely and fully for the purposes of the analysis of the actual and potential use of Prodigy Phase 3 with a possible analysis of the causes.

- All extracted data were cleaned and validated with other sources of information as well as for inherent consistency.

- The audit process did not involve any collection of patient identifiable information (only non-traceable clinical system internal codes were used). Neither the data contain any information about the GPs.

The raw data from the practice was initially imported into an MS Access 2000 database and then transferred to an SQL Server 2000 database at SCHIN for further analysis. At some stages of analysis STATA statistical package was used to facilitate data processing.

The process of data cleaning included the following stages:

· Prescription authorisations and notes associated with dormant problems were excluded from analysis.

· People with hypertension who are younger than 16 years old were excluded from the analysis because Prodigy does not support the treatment of hypertension in children. There was only 1 such patient. There were no people with angina younger than 16 years old.

· Only people with asthma who have had a respiratory drug within the last 4 years were included in the analysis. It was found that GPs often forget to change the status of asthma problems to dormant when they are no longer relevant to the patient. It was deemed reasonable to consider asthma records to be no longer current if the patient no longer takes any respiratory medication. It was found that over 22% of patients with a current problem of asthma had not been prescribed any respiratory medication within the last 4 years. This action left 1,163 patients with asthma from the total of 1,494.

3. Results

3.1. Number of patient records for the selected condition

The following table represents the number of patients with the current problem headings of hypertension, ischaemic heart disease (IHD) and asthma.

Hypertension		IHD, MI & Angina		Asthma				List size
Number of patients	% of list size	Number of patients	% of list size	Total current		After “cleaning”
				Number of patients	% of list size	Number of patients	% of list size
1373	13%	696	7%	1494	14%	1163	11%	10577

Table 1: Number of patients for the analysis

The proportion of patients with asthma record is more than 7% even after data cleaning, so there is no evidence of under-diagnosis.[BNC1] This value was suggested as a “rule of thumb” for clinical audit purposes in the work of Lakhani et al, 2000.

It was found that the practice uses IHD Read Code (G3…) for most of people with angina pectoris and myocardial infarction. Although not strictly wrong, such way of coding complicates audit[BNC2] . It is not possible to identify people who have had a prior myocardial infarction or angina pectoris. This data was not excluded from the analysis, but the analysis was limited to the criteria appropriate for all types of IHD (aspirin prophylaxis, prescribing of beta-blockers, monitoring).

Read Code	Problem term	Number of patients
G3...	Ischaemic heart disease	690
G30..	Acute myocardial infarction	3
G300.	Acute anterolateral infarction	2
G308.	Inferior myocard. infarct NOS	1
G30z.	Acute myocardial infarct. NOS	1
G31..	Other acute/subacute IHD
G311.	Preinfarction syndrome	2
G33..	Angina pectoris	9
G34..	Other chr.ischaemic heart dis.
G340.	Coronary atherosclerosis	2
G343.	Ischaemic cardiomyopathy	1
G37..	Cardiac syndrome X	1
G3z..	Ischaemic heart disease NOS	1

Table 2. Number of IHD patients by problem titles

3.2. Asthma audit results

This section provides a summary of results for the audit criteria of asthma. Every table represents a new criterion. The analysis for every criterion is based on a corresponding sample that is described in the first line of the table. The following line describes the actual criterion and provides both the number of patients with medical records that match the criterion and the corresponding percentage from the sample of the criterion.

1. Is it recorded that inhaler technique has been checked in the last year?

Total patients with asthma after data cleaning	1163
Inhaler technique monitoring recorded last year[1]	129 (11%)

Table 3. Recording of inhaler technique monitoring

It can be seen from the table above that the quality of recording of inhaler technique monitoring is quite poor: 89% of patients with asthma did not have a recent record of inhaler technique monitoring. It must be mentioned, however, that GPs may have their own special way of recording this action [BNC3] and certainly this is by no mean a valid estimate of the actual situation with inhaler technique monitoring. Prodigy Phase 3 would be able to assist in recording of such monitoring. The software will suggest that an inhaler technique monitoring is due for a patient that had no record of such within the last 6 months. If the user commits to this suggestion then a record of such monitoring is added to the patient record under the problem heading of asthma.

2. The number of children under the age of 4 years inclusive with an authorisation of beta₂ agonist using a breath-actuated device.

A breath-actuated device should not be prescribed to a small child, as it will be too difficult for the child to operate it. This criterion is not yet implemented in Prodigy Phase 3, but it will be included in the next update of the knowledge base.

Number of asthmatics 4 yrs inclusive with an authorisation of a beta₂ agonist (beta-adrenoceptor stimulant)	32
Number of children under the age of 4 years inclusive who are prescribed beta₂-agonists using a breath-actuated devices.	7 (24%)

Table 4. Breath-actuated devices for children under 4 years old

It can be seen that the practice has as many as 7 children under the age of 4 years who have been prescribed a breath-actuated device. The second release of Prodigy will have an additional functionality to “rule-out” breath-actuated devices for small children[BNC4] .

3. The number of children under the age of 4 years who are currently using a metered dose inhaler (MDI) who have NOT been prescribed a spacer device. The analysis is based on 4 years worth of medical records.

Number of children with asthma <= 4 yrs with MDI inhaler[2]	27
Number and percent of children with asthma <= 4 yrs with MDI inhaler and spacer record	24 (89%)
Percent of children with asthma <=4 yrs with MDI inhaler & NO record of prescription of spacer from total asthmatics <=4 yrs with MDI	11%

Table 5. Small children prescribed an MDI with no record of spacer device prescription

It can be seen that only 3 children under 4 years with an authorisation of an MDI did not have a record of a spacer. This could be attributed to various reasons, e.g. the spacer has been given in the hospital[BNC5] . It does not appear that there is a lot of room for improvement on this criterion.

4.The number of adults (older than 16 years) with a diagnosis of asthma with smoking status recorded.

The number of adults (> 16 years) with a current record of asthma	1100
The number of adults (> 16 years) with a current record of asthma with smoking status recorded within the last 4 years	1019 (93%)

Table 6. Recording of smoking status

Recording smoking status is of utmost importance for asthma monitoring. NICE asthma guidelines and many others actively promote recording of smoking status. It can be seen that the practice has a reasonably good quality (93%) of recording of smoking status for patients with asthma. Again it does not look as though Prodigy is likely to cause a noticeable improvement on this criterion.

6. The number of patients with a diagnosis of asthma on high daily doses of steroids[3] in a form of MDI who are prescribed these drugs with a spacer.

Dosage instruction is a user-editable field in the Torex clinical system and thus the data required manual coding. From 1,414 authorisations for the selected inhaled MDI steroids in a form of MDI prescribed for 440 patients it was possible to manually encode 1,382 (97.7%). This reduced the cohort to 433 (98.4%) of dose instructions. The bulk of the remaining 2.3% of authorisations had a synonym of “as directed” or “as required” entered in the patient record.

Patients with asthma on >=500mcg fluticasone a day	24
Above with a prescription of a spacer	6
Percent of those with a prescription of a spacer	25%
Patients with asthma on >=800mcg budesonide a day	3
Above with a prescription of a spacer	1
Percent of those with a prescription of a spacer	33%
Patients with asthma on >=1mg beclomethasone a day	42
Above with a prescription of a spacer	8
Percent of those with a prescription of a spacer	19%

Table 7. Patients with asthma on high dose steroids with a record of a spacer prescription

This criterion has shown that in general less than 40% of patients with asthma on high daily dose of steroid have a record of a spacer device. This analysis excluded patients with asthma who have been prescribed their steroids in any other form than MDI. The sample also excluded those who have been prescribed a steroid in a form of MDI together with a spacer as a single Read Code (e.g. “c61S.” for Beclomethasone dipropionate 250micrograms inhaler and compact spacer). Currently Prodigy cannot control prescription of spacers automatically. This is because the daily dose within the clinical computer system is a user-editable field, which cannot be used for making direct inferences. Prodigy Phase 3 does provide easily accessible “on-line” help that gives advice on prescribing spacers with high dose steroids.

Daily dose, mcg	100	200	400	600	800	1000	1200	1600	Total
Number of authorisations of beclomethasone	92	249	542	58	180	93	11	6	1231
Percent from total authorisations of beclomethasone	7.5%	20.2%	44.0%	4.7%	14.6%	7.6%	0.9%	0.5%	100.0%

Table 8. Number of authorisations of beclomethasone recorded in the practice for patients with asthma
Note: Columns with frequencies less than 5 are not shown

Table 8 shows the number of authorisations of beclomethasone, for different daily doses. There less than 2% of authorisations of daily dosages of beclomethasone of more than 1 mg. The most common daily dosage of beclomethasone is 400 mcg, which accounts for more than a half of all authorisations of beclomethasone.

Prodigy Phase 3 recommends adding a long-acting beta₂ agonist instead of stepping up a beclomethasone therapy when the patient is on a daily dose of 1mg of beclomethasone. Clearly the practice rarely prescribes high doses of beclomethasone leaving little room for improvement on this criterion.

3.3. Angina audit results

As described previously, the practice tends to use a read code of IHD (G3…) for all IHD – related conditions, e.g. myocardial infarction or angina. Although Prodigy Phase 3 does currently support only angina pectoris, it was decided to use all IHD patients for the purposes of the audit to increase the sample size. All the criteria included in this audit are equally valid both for angina pectoris and IHD patients. Throughout this section words “IHD patients” include patients with a Read code of ischaemic heart disease, angina pectoris and myocardial infarction.

1. Number of IHD patients with asthma who have been prescribed beta-blockers.

It was important to make sure that this analysis selects only those patients who were prescribed a beta-blocker after the diagnosis of asthma has been recorded. Otherwise there is no evidence that the clinician knew that the patient has asthma or is pre-disposed to asthma. Only asthmatics with a current respiratory medication were considered. Both restrictions have been taken into account when the figures for the table below where calculated.

Number of asthmatics on a beta-blocker within the last 4 years	62
Number of asthmatics on a beta-blocker within the last year	43
Total number of IHD patients with current asthma	98 (100%)
Number of IHD patients with asthma on a beta-blocker within the last year	16 (16%)
Number of IHD patients with asthma on a beta-blocker within the last 4 years	23 (23%)

Table 9. Number of patients with asthma having an authorisation of a beta-blocker added to their EPR

It can be seen from the table above that 42 patients in the practice have been prescribed beta-blockers despite having a record of asthma with a recent respiratory medication within the last year! This is quite a dramatic result given that this number only includes those patients with asthma who have had a recent respiratory medication and had beta-blockers prescribed after the diagnosis of asthma was recorded in their EPR.

It was found that the Torex System 6000 does not control for this contraindication and does not issue a warning when a user tries to prescribe a beta-blocker to a patient with a record of asthma[BNC6] . It does provide information about the contraindication under the “Info” tab in medication properties, but it takes few mouse clicks to get there and it is not apparent that such contraindication exists. Prodigy Phase 3 on the contrary actively helps to avoid prescribing beta-blockers to patients with an asthma record. If a patient has a record of asthma in the electronic patient record, Prodigy Phase 3 will either hide beta-blockers from a list of available options or will mark it as a “ruled-out” option providing quick “in-line” explanation together with evidence grade and detailed explanation of the reason that it is ruled out.

2. Aspirin prophylaxis

The following table shows the results of analysis of aspirin prophylaxis.

Sample	Number of patients	% from total IHD (N=696)
With an authorisation of aspirin	492	71%
With an authorisation of any antiplatelet drug (including aspirin)	521	75%
Adverse reaction to aspirin (presence of Read code: U6051 or TJ53.)	3	0%
Sensitivity for aspirin is recorded	6	1%
Discontinuation for aspirin is recorder other than "Added by mistake" or "Form/dosage Change"	59	8%
On aspirin or on any other antiplatelet drug or have a record of sensitivity or discontinuation for aspirin	589	85%

Table 10. Aspirin prophylaxis

Aspirin prophylaxis is difficult to audit because there is no standard way of recording it in the electronic patient record. It was found (see table above) that a large proportion of patients have authorisation records for aspirin and most of them (98%) are getting prescriptions for it. This is probably because people with angina are predominantly elderly and they often have exemption for their prescription charges.

Aspirin prophylaxis may also be recorded in a number of other ways – e.g. as a note of “Aspirin prophylaxis” (8B63.)[BNC7] or “Over the counter aspirin therapy” (8B3T.). The practice can also create their own coding for aspirin prophylaxis. This analysis focused only on actual prescriptions of aspirin. Thus real numbers of angina patients on aspirin prophylaxis may be even higher. It does not appear that Prodigy Phase 3 is likely to have significant impact on improving this situation, although there may be a room for improvement for aspirin prophylaxis. It is difficult to make a definitive conclusion as low dose aspirin can actually be obtained over the counter from pharmacies without prescription. As such a patient with angina may not necessarily have an authorisation of antiplatelet aspirin despite being on aspirin prophylaxis. Full audit of aspirin prophylaxis requires more detailed data, which was beyond the scope of this study.

3. Number of IHD patients prescribed a beta-blocker

IHD patients (excluding patients with a record of COPD[4], Asthma, Heart Failure, LVF or RVF)	477
IHD patients having an authorisation of beta-blocker (excluding patients with COPD or HF)	297 (62%)
Percent of IHD patients with no COPD or HF and not on beta-blocker	38%

Table 11. Number of patients with IHD on a beta-blockers

There are no fixed standards for the percentage of IHD patients that should be on beta-blockers, although it is strongly recommended (Chronic Heart Disease National Service Framework guideline and NICE guideline) unless there is a contraindication or intolerance. The number of IHD patients on beta-blockers appears to be appropriate. There may still be a room for improvement on this criterion, but it is difficult to estimate the extent of such an improvement (user compliance with guidance, patient compliance with therapy, ceiling limit, etc). There may be a number of good reasons why 38% of IHD patients did not have a record of beta-blocker for the past 4 years. It might be that a beta-blocker was unsuccessfully tried before and has been discontinued. It may also be the case that the angina is well controlled with a calcium channel blocker or a nitrate.

4. Number of patients with a record of smoking status, blood pressure, and lipid/cholesterol level.

It must be made clear at this point that these criteria are more to do with the quality of electronic patient records rather than directly with the quality of medical care.

The absence of a record does not necessarily imply that the measurement was not taken or the lab test was not ordered. It may be that the results were not recorded at all, or recorded in a non-standard way so that the audit query did not capture that record. Thus this part of analysis is more to do with the quality of data recording in practice clinical systems, rather than with the quality of care.

Total number of IHD patients	696
IHD patients with a record of BP within the last 4 years	694 (100%)
IHD patients having a record of BP within the last 12 months	668 (96%)
IHD patients having a record of smoking status within the last 4 years	686 (99%)
IHD patients having a record of smoking status within the last 12 months	525 (75%)
IHD patients having a record of either serum lipids (44O..) or serum cholesterol (44P..) within the last 4 years	651 (94%)
IHD patients having a record of either serum lipids (44O..) or serum cholesterol (44P..) within the last 12 months	576 (83%)

Table 12. Recording of blood pressure, smoking status and cholesterol level for patients with angina

It can be seen from the table that the practice have an exceptional quality of recording of patient information, measurements and lab results. Although here is still a small room for improvement in recording of recent data, it is difficult to tell if Prodigy Phase 3 is going to have a significant impact on these.

3.4. Hypertension audit results

Recommendations of British Hypertension Society (Williams et al., 1999) for blood pressure targets are presented in the following table:

Survey of care of asthma, IHD, and hypertension