1. HOSPITALS TODAY JK-PRACTITIONER
CLINICAL AUDIT
Sved Amin Tabish MD. FAMS. FlMSA. FRIPH
For a variety of reasons -professional, public and
political-most countries now seek to establish visible systems
for managing quality in health care. The motives, the settings
and the details may vary, but the general challenges facing
the enthusiasts are remarkably common.
‘Audit’ should include measurement of current practice,
and definition of explicit standards and implementing
changes which benefit patients; ‘clinical’ should imply
patient-centred and involving more than one clinical
profession (the experience of the patient is the unifying
factor).
Purpose
Decide why you want audit. Is it primarily for public
accountability, for management control, for contract
monitoring, or for professional self-development? Many
clinicians accept audit as an educational activity, led by the
profession but reported in general terms to managers. The
putpose of unidisciplinary audit has been stated as:
+ To improve the quality of patient care
+ To &rther professional education and training
+ To promote the effective use of resources.
Multidisciplinary audit concerns not only the clincial
practice within individual professions but also demonstrates
the contributions made by each and the organizational links
between them. Audit in many units thus reflects the clinical
directorate and health care team structure and includes a
further objective, ‘to define and monitor standards for the
provision and organisation of clincial services.’ Moving to
multidisciplinary clincial audit thus also implies a greater
involvement of managers, both providers and pruchasers.
Limitations
Fix the boundaries of audit. It relates to but is not the
same as several other activities which have different aims,
accountability or funding. In particular, audit should not be
confused with the following:
Research
This aims to identify general truths about health, disease
and the technology for intervention (what is the best
practice?); audit is about ensuring compliance with best
JK-Practitioner 2001; 8(4): 270 - 275
practice and a culture of constant evaluation and
improvement ofpatient care (do you apply the best practice?)
Resource Management
Good clinical practice is usually also efficient but
controlling resources is not the prime goal of audit. Even
so, resource management and audit share not only methods
and potentially f?uil.ful topics, but also the need for accurate,
complete and timely clinical data.
Contract Monitoring
Clinicians should be involved in setting and committed
to Wfilling quality clauses in service contracts, but data
generated by audit should not be assumed to be appropnate
for monitoring them, nor should audit staff be diverted to
collecting data solely for contract monitoring. Chnicians and
managers must agree on the responsibilities for collecting,
interpreting and exchanging clinical data for different
purposes.
Organization
Peer Group
For the purpose of audit, the appropriate ‘peer group’
depends on what is the topic under review. Clinical audit is
a method of formally questioning the interacting areas of
concern amongst different professions and parts of the
service. Its strength is that it takes the perspective of the
recipient, of the service and identifies the coordination of:
care. Staff should be reassured that clinical audit does not
signal the end of unidisciplinary audit. Exclusive meetings
by single professions or levels of seniority remain appropriate
for certain issues. To permit this flexibility, disciplines and
specialities need to share regular meeting times and to set a
programme of ,topics, may be six months or more in advance.
Structure
Define the local leaders and pathways. Plan for evolution
at a pace which is sensitive to the willingness of indlvldual
specialities, disciplines and units. Identify contact clinicians
in each and define how they relate to the committee structure,
to ‘quality assurance’ and to management.
Many units set a meeting once a month for medical audit,
and some have extended this to all clinical staff to enable
From the Department of Hospital Administration, SKIMS, Soura. Srinagar, Kashmir, India (Tabish)
Received May 2001 Accepted July 2001
Correspondence: Dr. Syed Amin Tabish MD, FAMS, FIMSA, FRIPH Sr, Consultant & Medical Administrator SKIMS Soura, Srinagar, Kashmir-
190011.
Vol. 8 No. 4, October - December 2001
2 7 0
2. -? 1’: 1 , .“...
JK-PRACTITIONER
their paticipation and to minimize disruption to services.
This requtres a clear management commitment not only to
audit in nursing, medical and the principal therapy
professions (clinical psychology, physiotherapy,.
occupational and speech therapy), but also in those such as
dietetics, pharmacy and otthoptics.
Policies I
Take stock of existing local agreements, for example, on
confidentiality, data validation, attendance records, reportmg
and sharing of information. Are they acceptable or adaptable
to all? :
A p p r o a c h e s
Principles I
Define the starting point. Are disciplines and team
members starting from the same point, for example, sharing
a level of understanding and definitions of terminology and
willing to acknowledge alternative viewpoints on quality,
to discuss professional overlaps, to recognize differing
models of patients care and and to acknowledge each other’s
skills? f j
Topic selection
Define priorities for audit. Define the basis on which
the team might set priorities for selecting audit topics, for
example, high volume, risk, cost or wide variation or an
area of local concern. The method of selecting topics should
incorporate all members of the group and be sensitive to the
legihniate interests of managers and the organization as a
whole. It should also be able to respond to relevant external
initiatives such as ‘consumer’ demands and national audit
proJects. i I
Methods
Identify available sources of advice and provide practical
guidance on appropriate methods. Define general
characteristics, for example, objective, quantified, standards-
based, repeatable. How is effective methodology to be
monitored and promoted?
Resources
Fix the amount of time to be devoted to audit. Allow for
duration and frequency of formal meetings as well as the
time set aside for their preparation and follow-up. In some
services, this has been assessed at half a day per month or 5
percent of clincial time.
Identify and develop sources of information for audit. Ln
particular:
I
a. Are clinical records accessible? Do they include
information on what patients are told, outcomes, etc?
b. Are statistics on case mix, complications, procedures and
treatments accurate, up-to-date and readily available to
clinicians for audit?
c. Are facilities available to all clinical staff to access
literature on clinical practice, and on audit, for example,
journals, newsletters, and annual reports?
Catalogue the uni-or multidisciplinary networks for
sharing methods, results and traming at regional and national
level. Are clinicians aware of these and have access to them?
Identify staff available to support chmcal audit. Clarify
their function, accountability and finnding. Who determines
the allocation to audit suppOrt time, and on what criteria?
Are the support staff sufficient in numbers and skill mix to
service the audit needs ofall clinical groups? Do Information
staffgive time to audit? Is there clerical and technical help?
Define responsibility and funding for training. This
should include audit staff, clinicians and managers.
Recognize the need for behavioural as well as technical
skills. What facilities are available locally, regionally or
nationally?
Identify the source and application ofexisting funds. How
is money provided for clinical audit, for example, nursing
and professions allied to medicine, hospital medical and
dental, primary health care, project grants, and any local
topup f?om general allocations.
Methods of Clinical Audit
Audit should involve the objective peer review ofpattems
of care, be sensitive to the expectations of patients and other
clinical disciplines and be based on scientific evidence of
good medical practice. To this end, cases for review can be
selected randomly or by pre-agreed criteria. Patterns of
practice, quantified where possible, can be compared wth
explicit guidelines and these guidelines, though agreed
locally, should be consistent with regional and. national
advice and with the body of scientific knowledge.
No single approach is universally applicable. Indeed, by
using a variety of methods, an audit group is more likely to
maintain flexibility, innovation and interest. Some methods
screen 100 per cent of cases for pre-defined events, which
are then subjected individually to peer review. Others-the
majority -focus on a limited sample of cases with a common
characteristic (such as a symptom, investigation, treahiient,
or complication).
Adverse patient Events
. ‘:” I_
Adverse patient events or ‘occurrence’ screening
involves the systematic identification and analysis of events
during a patient’s treatment which may indicate some lapse
in the quality of care. Screening criteria, aimed at
recognizing a particular occurrence, are defined by the
participating clinicians, but may be applied to individual
clinical records by non-medical assistants. For each criterion
there is a detailed definition of the occurrence-c 1t is intended
I
271
Vol. 8 IVO. 4, October - December 2001
3. JK-PRACTITIONER
to Idehtify, together with notes of known exceptions,
gmdellnes 011 use, examples, and details of the information
reqmred about each occurence. Fmdmgs are analysed by
the peer group to determlne causes and effects. This
information may be used to plan changes in patient
management.
It can be apphed across the hospital, or within a speciality
and is amenable to computer scanning if a comprehensive
database is available. It may be used for concurrent or
retrospechve screening of individual cases, or as a means of
identifying trends of suboptimal care. This technique is a
valuable tool in risk management But it is labour intensive
on :I conhnuous basis and its concentration on failures rather
than successes is not appealing to clinicians as a
comprehensive audit system.
Generic (whole hospital) criteria include the following:
+ Admission because of adverse results or complications
of out-patient management
+ ReadmIssion for complicatiuons or incomplete
management of problems on a previous admission.
+ Unplanned removal, injury or repair oforgan or structure
during surgery/invasive procedure.
+ Unplanned return to theatre.
+ Pnthologyihistology report varies significantly from
preoperativeiantetnortem diagnosis.
+ Cardiac or respiratory arrest.
+ Cardiovascular accidents or acute myocardial infarction
within 48 hours of surgical procedure, or pulmonary
etnbolus at any hme postoperatively.
+ Unexpected transfer to higher dependency unit.
+ Neurological deficit 011 discharge not present on
admission.
+ Unexpected death.
+ E,ldence of patient/family dissatisfaction.
TradItional morbidity and mortality meetings are
mechanisms for reviewing higIll>, selected adverse events
locally; confidential enquiries into maternal, perinatal, and
perioperative deaths, eclampsia and suicides while under
treatment are the equivalent at national level.
Clinic& Indicators
.J more comprehensive approach involves the monitoring
of routinely generated data within a speicality, in order to
identify excephons or trends which may merit detailed adhoc
re,iew. Ideally, each indcator would reflect an agreed policy
objective, be routinely and accurately captured and be
readily accessible.
Routine presentation of such data to clinicians, though
initially painful, does improve quality and usage. A general
model might Include measures, with relevant comparisons
01 er time or place, of
a. Workload: numerical description of cases, for example,
by type, source, site, age.
b. Access: measures of availability, for example, treatment
rates, waiting times, non-attendance rates, elective
admissions deferred by hospital.
c. Appropriateness: for example, operations out of hours,
readmissions, children in adult wards, patients admitted
after day case procedure.
d. Outcome: patients discharged home, in-patient deaths,
complications.
e. information: measures of quality (tnay be by sample
study), for example, cases with final diagnosis coded
within 1 week of discharge, records missing at time of
clinical visit/admission, concordance of parallel data
systems.
f. Efficiency: clinic/theatre sessions cancelled, length of
stay (selected ddagnosesiprocedures), new/old patient
iatio, pharmac cost per ward/clinic, investigations
requested, day cases as percentage of all planned
procedures.
Most of these are available from existing data but need
to be collected and refined.
Topic Review
Analysis of an agreed topic may be carried out by
prospective study or by retrospective analysis of patient
records. The latter often reveals inadequacies in recording
but is a valuable exercise in providing a baseline for
comparison in a later, prospective study. Both approaches
involve a systematic review of a large enough sample of
similar cases in order to identify, quantify and con pare local
patterns of practice.
The concept is applicable to surgical and non-surgical
specialities, is objective, yields quantitative data and is
repeatable but is very labour intensive, For this reason, the
clinicians may define explicit criteria by which an assistant
can extract key information from large numbers of medical
records. The steps in this ‘criterion-based’ audit are
represented in Table 80.1.
@bJa 80.1: Criterion-based Audit
Clinicians Audit Assislanl
Steps
1. Define topic.
2 Define criteria for analysis
3 Defrne sample SIZB, dates, etc.
4.
5
6.
7.
8. Discuss findings and exceptions.
9. Define action and fhre policy.
10 Repeat audit agarnsi criieria.
fdentify case: for analysis.
Frnd indivtdual medrcal records
Abstract data against criteria.
Tabulate aggregated data.
The topic should be of general interest and of high
Vol. 8 No. 4, October - December 2001
272
4. I . .I
JK-PRACTITIONER
volume, risk of cost. Or it should be a subject of wide local
variation or other concern. It may be a symptom, diagnosis,
investigation, treatment, outcome or just a clinical problem.
The snnpler and more common the topic, the easier it is to
collect data and repeat the audit before junior staff initially
involved in the work move on.
Discusuon on the criteria and sample details should
involve the audit assistant to minimize later problems of
interpretation. Typically, the criteria should form abou a
dozen self-explanatory queshons, with a numerical or a yes/
no answer which can be gleaned from the record.
The purpose of these criteria is not to act as a protocol
for clinical management but to establish current patterns
and to help identify records which may merit individual
review. Local groups may develop different ideas and these
may represent the lowest common denominator on which
&,reement is achievable. For a particular inpatient diagnosis,
these may be under headings such as:
1 !
+ Referral: for example, delay before seen by junior/
consultant.
+ History for example, specified symptom(s) recorded.
+ Examination: for example, specified sign(s) recorded.
+ Investigation: for example, specified test(s) reported.
+ Treatment: for example, specified drugs (type, dosage,
route( given.
+ Follow-up: for example, discharged asymptomatic within
agreed time from ward/from clinic.
+ Communication: for example, summary to General
Practitioner within ‘n’ days.
It is wise at this stage to test the criteria on a very small
sample of records, say five, to check that there is a realistic
chance of finding the required information.
The assistant then has the often challenging task of
finding an accurate, up to date, listing of patients who had
the specified diagnosis and were treated by the doctors
concerned within the defined time scale. The next task is to
find the records themselves, and not to bias the sample by
giving up on the ones that are missing.
Abstracting of the records, once found, can take I O-30
minutes per case, depending on the number and complexity
of the criteria. Multiplied by, say, twenty cases for each of
three or four firms involved, this occupies many hours of
work.
The results have to be tabulated and presented (ideally
with the help of a computer spreadsheet and graphics
package) to the chnicians. They must satisfy themselves
that the findings are accurate.
Discussion tends to turn to the choice of criteria to the
interpretation of results and to agreement on action.
The audit can be repeated, using the same or modified
criteria after an agreed hme in order to measure the impact
of the changes on patient care.
Medical Records
Clinical guidelines for minimum standards of records
have been issued by the Royal College of Surgeons, England.
Checklists for evaluating records were included in the first
audit report of the Royal College of Physicians, England.
Auditing the records of every patient who had general
anaesthesia should be done on the basis of the criteria shown
in the box.
Records correctly field
Percentage meet
Anaesthetic record present.
Patient name on sheet
Pre-operative assessment recorded
Operation date recorded
Operation type recorded
Anaesthetic technique specified
Blood pressure recorded
Oximeter reading recorded
Name of anaesthetist legible on record
Therapeutics
Prescribing
Audit of drug usage is valuable both clinically and
economically, and data (on choice, route, dose, and duration)
are generally well recorded. Such audit also highlights some
of the problems of implementing and maintaining change in
clinical practice.
Some studies of medications focus on choice of drug,
for example, thrombolytics; appropriateness; route of
administration; and serum monitoring.
Other forms of therapy also lend themselves to audit and
these often involve other clinical specialities and disciplines.
Non-surgical uses in the hospital could also be reviewed.
Although primarily relevant to dermatology, the use of PWA
treatment illustrates a general model for the discussion of
appropriateness which might precede audit of any tlierapy,
in terms of criteria for entry, process and exit.
Entry ‘i.#l .*;. ‘;;:‘.I’.: ._ -; ,,. .
On what criteria ae patients selected/refbsed?
4 diagnosis? ‘,.’
+ disability (physical/social)? .:1:.-.. ::c! ‘:. ‘:
+ extent of disease?
.‘- t
4 age of patient? . ‘.
+ failure of first line therapy? .
+ ability to attend for regular treatments?
‘_ -
‘,
Process ’ ‘;L< .;4 I
How should treatment be conducted?
+ what machine capacity / power is appropriate?
+ what staff supervise therapy? _.
+ safety/calibration procedures?. ‘: ’ i
Vol. 8 No. 4, October - December 2001
273
L
5. JK-PRACTITIONER
+ number of exposures/frequency?
Exit
When should treatment cease?
+ percentage improvement?
+ failure to respond?
+ side effects?
The calibration and safety of equipment used by doctors
IS central to the overall quality of medical care.
Diagnostic Investigations
Radiology
Audit of the use of diagnostic radiology is indicated by
evidence of clinically unproductive ‘routines’ in many
sepecialities,, by potential savings in inconvenience to
patients staff time and hospital costs, and by the growing
concern over exposure to medical radiation.
The Joint Working Party of the Royal College of
Radiologists, London and the National radiological
Protection Board, London estimated that the current
population dosage could be halved without detriment to
patient care. The remaining half would still be three times
greater than the total of all other man-made sources of
populationradiation. Someofthisreductioncouldcomefiom
various means of optimizing dosages per examination, but
the greater contribution would be in avoiding unnecessary
X-rays. At least 20 per cent of X-ray examinations currently
carried out in the UK are clinically unhelpful in the sense
that the probability of obtaining information useful for
patient management is extremely low.
The Royal College of Radiologists, London
recommended a ‘temporary norm’, based on the lowest rate
found, of I2 per cent.
The guidelines may easily be developed into explicit
criteria for the audit of current request patrems.
Post-Mortem Examination
Autopsy findings are consistently reported to differ
substantially from antemortem diagnoses in 10 to 25 percent
of cases-indeed the ‘ultimate audit.’
Random Case Review
Cases may be selected randomly or by a predetermined
system from among the general workload for critical review
by doctors not previously involved in the clinical
management. Findings can then be presented to the audit
group for discussion. Checklists may be used to ensure that
each case is reviewed according to an agreed system of
questions.
The method is easy to use, requiring minimal resources
and no numerical data. Many groups find it a useful
introduction to more structured discussion of clinical
management. But it requires considerable time from the
reviewers prior to the meeting. MaJor issues may be missed
through the randomness of selection. Discussion tends to
be subjective and criticism often centres on the inadequacy
of the record and thus the junior staff.
In many hospitals, one or two audit meetings a year are
given to random reviews as a change from more focussed
analysis of single issues.
Patient Satisfaction
The assessment of the views of patients and their relatives
must be incorporated into the judgments which doctors apply
to their work. Formal surveys of patients’ lews aregneerally
included in the overall quality improvement programmes of
hospitals, rather than in medical audit. But audit may
scrutinize evidence of thorough history-taking treatment
plans agreed with patients, informaiton given, perceived
value of treatment and outcome measurements expressed in
the patients’ own terms.
The honest views ofpatients anddoctors on the objectives
of medical intervention do not always tally. For example,
cancer patients may choose less interventional therapy m
preference to higher chance of 5-year survival.
Complaints leading to litigation are increasing and
analysis of the Icauses can provide guidance for avoiding
similar events III the future. Failure to communicate, explain
or apologize can add substantially to even a relatively minor
mishap and beccme a prime reason why patients or relaltives
pursue legal redress. Clinical and admmistrahve lessons can
also be gleaned fromthe experience of the denfence
organizations, and provide material for audit.
Comparative Audit
The confidential pooling of aggregated data, using
standard definitions and formats, allows Individual doctors
to receive feedback on their own performance compared with
others working in the same field. This can be simplified
with standardized data systems, but the method of production
of the statistics is less important than their accuracy. The
most reliable data are those collected by individual clinicians.
Many examples of the use of comparative data are found in
surgery but the principle has been applied wthln health
regions to obstetrics and medical specialities.
Further Reading:-
1. Bennett J. and Shaw C.D. (1987). ‘Guidance on what should
be in the clincial medical recoirds for the Bri$ton health
district’, Med. Rec. Healthcare Info.. 28: 103-10.
2. Bennett J. and W&he K. (1991). ‘Occurrence screening as a
method ofaudit’, Br. Med. J., 300;1248-51.
3. Donaldson L.J. and Hill P.M. (1991). ‘The domiciliary
consultation service: time to take stock’. Br. Med. J., 302:449-
51.
4. Feely J. et al. (1989). ‘Hosptial fonmdaries: need for contmuous
intervention,’ Br. Med. J.. 300:28-30.
Vol. 8 No. 4, October - December 2001
274
6. JK-PRACTITIONER
5. Gau G. (1977). ‘The ultimate audit’. Br. Med. J., 1:1580-l.
6. Health D.A. (1981). ‘Medical audit in general medicine’. J.
Roy. ~011. Phys. Lond., 15:197-9.
7. Leape L.L., BrennanT.A,and LainiN. etal. (1991). ‘Thenature
of adverse events in hospitalised patients’, N. engl. hled. J.,
324377-84.
8. Massanari R.M , Wolkerson K.. and StreedS. (1990).
“concurrent surveillance for adverse outcomes of medical care’,
in Quality in health care. Proceedings of conference of
lntematronal Society for Quality Assurance, ISQA Stockholm.
9. morgan D.J.r., Sutters C.A.. and Pugh S. (1991). ‘Medrcal audit
and formulary management: a policy for rational use of
thrombolytic drugs’. Postgad. Med. J., 67: 165-9.
IO. Roberts C.J., Fowkes F.G.R.. Ennis W.P., and Mitchell M.
( 1983). ‘Possible impact of audit on chest X-ray requests from
surgical wards’. Lancet, ii:P446-8.
11. Royal College of Physicians (1989). Medical audit: a first
report-what, why and how? RCP, London.
12. Royal College of Radiologsts (1989). making the best use of
department of radiology. PCR, London.
13. Royal College of Radiologists (1989). ‘Preoperative chest
I .,
radiolokv’, Lancet, ii:83-6.
14. Royal College ofradiohogists, National Radiologjcal Protection
Board (1990). Patient dose reduction in diagnostic radiolog,
NRPB, Didcot.
15. Royal College of Surgeons (England) (1990). Guidelines for
clinicians on medical records and notes, RCS. London.
16. Russell G.A. and Berry P.J. (1989). ‘Postmortem audit in a
paediatric cardiology unit’, J. Clin. Pathol.. 42:912-S.
17. Sandler G. (1984). ‘Do emergency tests help in the management
of acute medical admissions?‘, Br. Med. J.. 289:973-7
18. Shaw C.D. (1990). ‘Criterion based audit’, Br. Med. J., 300,
649-5 1.
19. Shaw C.D. (1992). Speciality Medical Audit, King’s Fund
Centre, London,
20. Simpson N.B. (1990). Medical audit for dermatologists: a
practical &aide, British Association of Dermatologists. Glasgow.
21. Swindell P.J., reeves D.S.. Bullock D.W.. Davies D.J.. and
Spence C.E. (1983). ‘Audits of antibiotic prescribing in a Briston
Hospital’, Br. Med. J.. 286: 118-22.
22. Young D.W. (1988). “Improving laboratory usage: a review’,
Postgard. Med. J.. 64:283-9.
_.
I .
Vol. 8 No. 4, October - December 2OOI
2 7 5