Norway

Dr T. Carlsen, Dr A. Stavdal

Contact Person:
Dr Tor Carlsen
Email tor.carlsen@klosterhagen.no
January 25. 2005

Problem
There is a need to develop effective tools for implementing clinical guidelines to promote effective learning methods in Continuous Medical Education (CME) for general practice.

Design
SATS is a quality cycle method for peer groups. SATS uses reports of quality indicators generated from electronic patient records (EPR).  
Background and setting
The SATS trial  (1996-99)  was conducted by a national project group of general practitioners . All GPs in Norway received information about the project with an invitation to take part and receive CME credits.

Key measures for improvement
Quality indicators were developed in four clinical areas in order to monitor clinical processes and outcomes.  Participants’ views of project participation were assessed using questionnaires. An independent qualitative study of participant response using focus groups was also performed.

Strategy for change
SATS combined the use of evidence-based indicators and specialised computer software.   “Pop up” forms helped collect data and provided on-screen support during consultations. Reports were used to stimulate systematic reflection on variation, the need for change and ways of changing clinical behaviour.  Data was kept anonymous outside the groups.

Effects of change
Most participants reported improvement as measured by one or more indicators. Those with the lowest scores showed the most substantial improvement.

Lessons learnt and next steps
The SATS method was well accepted by the participants. We must   simplify all elements of  the method; selection of indicators, the user interface of the computer modules, and the aids for the quality cycles. There is a need for stronger support of the group leaders. A large scale follow up of SATS is being launched (2003). 

Brief description of context
Several quality improvement methods have been introduced into Norwegian general practice over the past twenty years, such as audit, practice visits, case analyses, evaluation of referrals, consultation video analysis and the use of simulated patients. The methods have all been implemented as part of the Continuous Medical Education (CME) system, but few of them have achieved widespread use.  
Over 90% of general practice records in Norway are computerised but the resultant outcome of enhanced quality improvement in patient care has been difficult to achieve.
Substantial resources have been used to establish national guidelines for chronic conditions such as diabetes mellitus, hypertension and asthma.  There has been a lack of methods to implement guidelines in clinical practice, and their resulting impact has been limited.

Outline of problem
Traditional methods of CME often give a low yield in terms of inducing change in clinical behaviour1.  Modern learning theory emphasises that self-monitoring of continuous education improves outcome, particularly when learning is based on experience of the learner’s own practice2.  The influence of colleagues and collaborators may enhance learning3. Peer group learning with reflection on the practice of the participants may be an effective way of changing clinical performance 4,5.  
The aim of SATS was to develop a feasible, measurable and effective method of change in clinical performance. 
The SATS project  (1995-98) was carried out by a group of five GP’s, one from each region of the country.  The medical association covered study expenses through its Quality Assurance Fund.

Key measures for improvement
The group established indicators following a literature review and discussions with GPs and academic resource persons. Four sets of indicators were established, covering diabetes care, diagnosis and treatment of sore throat, migraine, and the clinical use of laboratory services. Examples of these indicators are presented in table 1. 

Objective  Indicator
Correct diagnosis  Is the streptococcal antigen test positive or are clinical criteria fulfilled when antibiotics have been given? 
Correct choice of antibiotic  When an antibiotic other than Penicillin V has been given – is there allergy or recurrence?
Dosage and treatment as recommended  Have antibiotics been given for at least ten days? Is the dose less than 1200 mg/day?

Objective		Indicator
Correct diagnosis		Is the streptococcal antigen test positive or are clinical criteria fulfilled when antibiotics have been given?
Correct choice of antibiotic		When an antibiotic other than Penicillin V has been given – is there allergy or recurrence?
Dosage and treatment as recommended		Have antibiotics been given for at least ten days? Is the dose less than 1200 mg/day?

Table 1. Indicators – the diagnosis and treatment of sore throat


Process of gathering information
Electronic tools were integrated into the electronic patient record in order to simplify the registration and retrieval of data. Short pop-up forms were completed on the screen during the consultation. Each GP received a diskette with a feedback report containing a survey of his/her practice. A typical report would include 10-20 patients with a defined clinical problem.


Strategy for change
Peer groups were established to review individual members’ reports.  The groups began by discussing variation among the members and then proceeded to identify needs and solutions for change in practice. Typical proposals for change included updating of practice protocols based on changes in knowledge, education and delegation of responsibility to assistant personnel (nurses, secretaries), and the purchase of new equipment.  Project booklets presented questions and proposals, while the solutions were left to the participating GPs.


Analysis and interpretation
Anonymous copies of the computer reports were used to develop performance data. Questionnaires gave information about the participants, their motivation for taking part, and their experience of group work. GPs could compare the degree to which each indicator was met before and after the intervention. As the number of patients was usually low, and only two counts (before-after) were taken, it would take a substantial change to ascertain that an improvement had taken place for the individual GP. There was no control group data, which could correct for general trends independent of the project. Analysis was done on the whole group of patients, which may hide cluster effects.


Effects of change
All Norwegian GPs were invited by to participate in the project by letter. 53 groups registered their interest, of which 35 commenced and 30 completed the process, with a total of 180 physicians.  The participants had above average experience in general practice6. Most indicated a positive attitude towards group work as the main reason for participating (Table 2).

	Number	(%)
Interest in quality work	135	(73)
Credits from the group work	122	(69)
Positive attitude to the method in the group	116	(66)
Interesting design	124	(69)
Positive attitude to peer group work	170	(89)

Table 2. The participants’ motivation to take part, at the outset. (180  respondents)

The comparison of reports from the first and the second registration shows improvement in those indicators that describe the working process (Table 3).

Criterion	First registration N = 747 patients		Second registration N = 728 patients		P-value
	Number	(%)	Number	%
HbA1C measured during last year	690	(92)	705	97	<0,01
Urine Microalbumin has been investigated	207	(28)	275	(38)	<0,01
Blood pressure has been measured	647	(87)	664	(91)	<0,01
Serum cholesterol has been investigated	438	(59)	455	(63)	0,142
Serum HDL-cholesterol has been investigated	270	(36)	375	(52)	<0,01
Serum triglycerides have been investigated	307	(40)	365	(49)	<0,01
Smoking habits have been documented	376	(50)	664	(88)	<0,01
The fundi have been examined	493	(66)	560	(77)	<0,01
The feet have been examined	294	(39)	492	(68)	<0,01
The patient monitors their own blood sugar	358	(50)	373	(54)	0,204

Table  3. Improvement in clinical indicators in nine diabetes groups with 57 participants, (percent fulfillment of quality criteria).

Patients with diabetes were well monitored with respect to glycaemic control, risk factors and complications. Prescription of antibiotics in sore throat changed, as recommended, towards more use of Penicillin V, lower doses and a longer treatment period (Table 4).  The physicians excluded fewer cases with low haemoglobin from follow up.

Criterion	First measurement		Second measurement		P-value
	Number	(%)	Number	(%)
Antibiotics given despite negative antigen test	66 of 43	(46)	43 of 31	(33)	0,033
Allergy or recurrence when antibiotic other than Penicillin V has been given	26 of 10	(24)	34 of 54	(63)	<0,01
Proportion of antibiotics which are broad spectrum	110 of 398	(28)	54 of 483	(11)	<0,01
Antibiotics given for less than ten days      (low % desirable)	135 of 398	(34)	11 of 483	(2)	<0,01
Daily Penicillin V dosage above 1200 mg (low % is desirable)	231 of 363	(64)	117 of 430	(27)	<0,01

Table 4  Improvement in the use of antibiotics with sore throat. (52 physicians, 430 patients)

Almost all GPs improved some aspect of clinical performance during the process, and those with the least favourable commencement data improved the most. Table 5 demonstrates that the 20% who had the lowest fulfilment of criteria for correct laboratory use in the first registration, approached the average in the second registration. The ”poorest fifth” of the GPs showed more than half of the registered improvement.

Criterion	First measurement				Second measurement
	All physicians		Poorest fifth		All physicians		Poorest fifth at first measurement
	Number	(%)	Number	%	Number	%	Number	%
Ferritin measured during pregnancy	1097 of 1428	(77)	81 of 228	(36)	1 137 of 1301	(87)	222 of271	(82)
HbA1c measured during previous year (diabetics)	1304 of 1559	(84)	69 of 128	(54)	1 074 of 1162	(92)	120 of 148	(81)
Abnormal Hemoglobin followed up by GP	601 of 737	(82)	96 of 169	(57)	673 of 743	(91)	160 of184	(87)

Table 5. Improved use of laboratory tests in 11 groups with 66 physicians.  Comparison of the fifth with the poorest initial results with the whole group of participants

74% of the participants said that they wanted to take part in new topics with the same group, and 81% would recommend SATS to colleagues.  Among those who completed the cycle, 61 % said that their own practice had been improved. 24% thought that the activity had improved practice in other areas. The response was most positive in the groups who had worked with diabetes and least positive in the sore throat groups. The data was analysed using the Epi Info statistics programme. A chi square test was used to compare data before and after the intervention.


Effects of change
The study indicates improvement in most indicators of the clinical work process. The goal of quality improvement is to improve health services to as many patients as possible. Improvements with those physicians who perform at the lowest level, will contribute most strongly to this aim.  Traditionally, CME is reported to give the most benefit to the participant who is the most interested beforehand 7.  Both the diabetes and the laboratory groups seemed to show the opposite trend.  We think that this is due to the stimulating effect of the group discussions. If this can be confirmed in later investigations, it should be important in the further development of methods of improvement in primary care.
We asked the participants to evaluate quality related to structure, process and outcome, in line with Donabedian 8.  The participants showed little interest in continued recording of information about routines and resources.  The majority of diabetes participants did not carry out collection of data on   patient satisfaction. We think that this was because they thought the perceived benefit was not in proportion to the input required.  The recording of clinical outcome, for example the number of recurrences of sore throat, metabolic control in diabetics or control of TSH values in patients with hypothyroidism, did not provide much benefit. This may be due to the small number of patients per individual GP, or incompleteness of follow up data. 
The participants valued the opportunity of comparison of their own practice with that of others as a good starting-point for learning. The group discussions provided support and constructive criticism, which are both important for learning benefit 9. The participants discussed what they were actually doing with patients, rather than what they assumed they were doing.
The strength of the method seems to be that it is linked to personal data, woven into clinical everyday work, discussed with colleagues and is relatively simple to use.  The participants provide the data themselves, independent of external support. This gives flexibility. The method presupposes full openness about data inside the group, but gives protection from external observation. It is seen as crucial that the participants “own” their data and remain confident that the aim is professional development, not external control. The social ambiance of the peer group was also seen as important 9.


The lessons learnt
The SATS project combined known strategies for CME with the aim of providing better services for patients and effective learning for the GPs. This was partly successful. Despite practical barriers the trial showed that this is attainable, and that that GPs may be interested in using this type of tool.
The development of SATS was laborious because the concept was new and the technical part was complicated. The participants needed a lot of time for practical preparation.  We assume that the participants were more motivated for and more experienced with quality development work than the average GP. However, the results are seen as useful for the continued development of the method, in terms of resource estimation, recruitment and practical schemes. 
There were a number of problems related to computer hardware and software.  Some groups reported lack of response upon submitting the material, and some thought that the tools were too complicated to use 9. 
 

The next steps
We have not looked at the long-term performance data. Other authors have found that reinforcement of learning is necessary to maintain the benefit of a pedagogical intervention 10.In a future development of the SATS method we need to investigate the prerequisites for a long-term effect.  The need for support of group leadership should also be systematically examined.
We are currently (2003) launching the second generation of SATS, called “NOKLUS-klinikk”with emphasis on diabetes and the clinical use of laboratory tests.  Software, administration of data and presentation of clinical problems have been updated and simplified. The aim is to encourage SATS participation within the CME system and the reimbursement system. The advent of a national digital health net will enhance the logistics and the marketing of the project.
 
 
References

1. Davis DA, Thomson MA, Oxman AD, Haynes RB. Changing physician performance: a systematic review of the effect of continuing medical education strategies. JAMA 1997; 274:700-5.
2. Cantillon P, Jones R. Does continuing medical education in general practice make a difference? BMJ 1999; 318: 1276-9.
3.Hjortdahl P. Hvorfor og hvordan endre legers kliniske atferd? (Why and how change the clinical performance of doctors?)Tidsskr Nor Lægeforen 1994; 114: 2372-3.
4. Pendleton D, Hasler J. Professional development in general practice. Oxford: Oxford University Press, 1996.
5. Pedersen OB, Prestegaard K, Hjortdahl P, Holm HA. Livslang læring – hvordan lærer leger?( How do doctors learn) Tidsskr Nor Lægeforen 1996; 116: 2684-8.
6. http//www.legeforeningen.no, legestatistikk, 19.3.1999.
7. Sibley JC, Sackett DL, Neufeld V, Gerrard B, Rudnick KV, Fraser W. A randomized trial of continuing medical education. N Engl J Med 1982; 302: 511-5
8. Donabedian A. Evaluating the quality of medical care. Med Care 1968; 6: 181-202.
9. Rørtveit G, Schei E. Qualitative evaluation of SATS. Bør man satse på SATS? Tidsskr Nor Lægeforen 1999; 119: 2689-72. 
10. Rutz W, Von Knorring L, Walinder J. Long-term effects of an educational program for general practitioners given by the Swedish committee for the prevention and treatment of depression. Acta Psychiatria Scand 1991; 85: 83-8.
 

 

 

Allmennmedisinsk Kvalitetsutvalg