Variation in purchasing for the invasive management of coronary heartdisease

Public Health (1996) 110, 13-16 © The Societyof Public Health. 1996

Variation in purchasing for the invasive management of coronary heart disease DJ Gunnell and I Harvey

Department of Social Medicine, Canynge Hall, Whiteladies Road, Bristol, BS8 2PR This paper reports the results of a postal questionnaire survey of Directors of Public Health in all health authorities in the United Kingdom (as at March 1994). Our aim was to examine variations in the purchasing of coronary artery bypass grafting, percutaneous transluminal coronary angioplasty and coronary angiography. Information on planned service developments in cardiology was also sought. The response rate was 62%. The mean rate of CABG was 374 per million total population (range 162-710); PTCA 183 (range: 18-648); and coronary angiography 1,010 (range 581-2,334). The mean ratio of invasive treatment to angiography was 1 : 2. Variations in provision were not related to mortality from coronary heart disease or the availability of a local provider. Those districts purchasing higher levels of CABG tended to purchase higher levels of PTCA (Spearman's r = 0.52). Observed variations in purchasing of invasive treatments and investigation for coronary heart disease do .not relate to population 'need' as defined by mortality rates from CHD. The greatest variations are seen in the purchasing of PTCA, an intervention whose place in the management of CHD is as yet not fully defined. Consensus guidelines on the appropriate use of these interventions and on population needs are required. Key words: coronary heart disease, coronary artery bypass grafting, percutaneous transluminal coronary angioplasty, angiography, purchasing, variations.

Introduction Purchasing decisions should as far as possible be based upon evidence of population needs for health care interventions. ~ in particular the treatments and investigations purchased should be of proven clinical effectiveness. In the field of cardiac surgery there is uncertainty over the a p p r o p r i a t e balance in provision of percutaneous transluminal coronary angioplasty (PTCA) and c o r o n a r y artery bypass grafting (CABG). A n u m b e r of current randomised controlled trials are comparing the cost effectiveness of P T C A and CABG. However only short term follow-up results are yet available. 2,3 In the U K a recent review of C A B G and P T C A provision in four health regions d e m o n s t r a t e d m a r k e d inter-regional and inter-district variation in the provision of these invasive cardiac interventions, 4 Setting national target rates for C A B G and P T C A - such as that of 300 C A B G per million p o p u l a t i o n resulting from a 1984 consensus conference - m a y help to decrease the observed variation. 5 This target was a d o p t e d by the D e p a r t m e n t of Health. M o r e recently the Joint Cardiology C o m m i t t e e of the Royal College of Physicians and the Royal College of Surgeons has r e c o m m e n d e d that rates for C A B G of 400-500 per miltion and for P T C A of 300 per million would be appropriate. 6 These targets, unrelated as they are to p o p u l a t i o n prevalence of c o r o n a r y heart disease ( C H D ) and age-structure, should ideally be tailored to the specific epidemiological circumstances of individual districts. Progress towards such targets in a given p o p u l a t i o n m a y be m a d e less predictable by the purchasing decisions of G P fund holders and monitoring of progress is m a d e m o r e difficult by the unavailability of data a b o u t activity in the private sector. Correspondence to: Dr David J GunneH, Lecturer in Epidemiology and Public Health Medicine, Department of Social Medicine, Canynge Halt, Whiteladies Road, Bristol, BS8 2PR. Accepted 12th July 1995.

The a p p o i n t m e n t of increased numbers of fully trained cardiologists to district general h o s p i t a l s and the appreciation of the effectiveness of C A B G a n d P T C A has led to increases b o t h in the total n u m b e r of cardiac catheterisations performed and in the n u m b e r of District General Hospital based cardiac catheterisation laboratories. Although such centres are m o r e accessible for patients, it has been suggested that their ready avaiiability m a y lead to increased inappropriate use and without the b a c k up of on-site cardiac surgery facilities. Against this b a c k g r o u n d of uncertainty a b o u t the empirical base for rational purchasing and delivery of services, a postal questionnaire survery of commissioning agencies in the United K i n g d o m was undertaken. The principal objective was to determine current purchasing strategies and p r o p o s e d service developments and so d o c u m e n t the extent of variation in purchaser's behaviour.

Methods A postal questionnaire was sent to the 169 health authorities in the United K i n g d o m in M a r c h 1994. N o n - r e s p o n d i n g H e a l t h Authorities were sent a reminder with a shorter version of the questionnaire after four months. I n f o r m a t i o n was sought on the n u m b e r of CABGs, P T C A s and coronary a n g i o g r a m s contracted for in 1993/4. I n f o r m a t i o n was also sought on the n u m b e r of these procedures purchased by fundholding general practitioners. These figures were combined to give a crude rate of purchased procedures. Where contracts did not specify specific n u m b e r s of C A B G , P T C A and c o r o n a r y angiography, actual numbers of procedures performed were used as a p r o x y measure for this. I n f o r m a t i o n was also sought on whether providers within the district u n d e r t o o k C A B G , P T C A or c o r o n a r y a n g i o g r a p h y and on whether there were any plans for the development of these services.

14

Invasive management CDH-purchasing DJ Gunnell and I Harvay

D a t a from the 1993 Public Health C o m m o n D a t a Set 7 were used to assess representativeness of the responding English districts. A weighted age standardised mortality rate for all responding districts was calculated and the value for this c o m p a r e d with that for England as a whole. Standardised mortality rates for C H D in those aged less than 65 years and those aged 65-74 were used as a proxy measure of 'need'. Spearman's rank correlation coefficients were calculated to examine the relationship between S M R for C H D and rates of invasive investigation and treatment of C H D . This relationship was only examined for the Health Authorities of England as similar data for other districts of the U K were not readily available. Where districts were only able to report contract data for 1992/3 or 1994/5; these data were used only in analyses of the ratio of C A B G and P T C A to coronary angiography. In one region (population 5.2 million) information was returned for the whole region rather than from individual districts. In this region contracting was undertaken on a regional rather than a district basis. This was therefore treated as a single data point.

Results

Response rate and representativeness Responses were received from 105 health authorities (62% of those sent the questionnaire), covering a p o p u lation of 37 million. Some of these authorities had recently merged. At least three health districts from each health region c o m p l e t e d the questionnaire and responding health authorities covered 67% of the p o p u lation of England; 41% of the population of Wales; 66% of the population of Scotland and 43% of the p o p u l a t i o n of N o r t h e r n Ireland. Some districts were unable to provide data for 1993/4; data for this year were reported by districts covering a total population of 27,190,00 (47% of the population of the UK). A quarter of respondents used block contracts for cardiac services. These authorities reported actual numbers of procedures performed on their district's residents. D a t a from these districts have been included in all analyses. Their exclusion did not effect any of the results. The standardised mortality rate from C H D a m o n g s t the responding English districts was 54.0 per 100,000 for those aged < 6 5 and 849.4 per 100,000 a m o n g s t those aged 65-74. These compare with figures of 52.7 and 838.4 respectively, for England as a whole.

Rates of CABG, P T C A and coronary angiography Table I shows the mean, median and range of rates for CABG, P T C A and c o r o n a r y angiography specified in the responding districts contracts for 1993/4. D a t a were available for districts covering a population of 27,190,000, approximately half the population of the U K . The m e a n ratio of invasive treatments (CABG and PTCA) to coronary angiograms purchased was 1:2 (median 1 : 2.1, range 1 : 1-3.6). The m e a n rate of C A B G in those districts with a local provider of cardiac surgery was 372 per million and for other districts the rate was 365 per million. F o r P T C A the reported rates were 170 p e r million and 190 per million, respectively. F o r angiography these figures were 884 and 925 per million, respectively. Forty-five of the responding districts had set target levels of CABG. In some areas regional targets were used, whilst other districts had developed their own local targets. N o relationship was found between district targets for C A B G and either standardised mortality rates for C H D or current rates of CABG. Three districts from two different regions reported targets for C A B G of 500 per million to be achieved by the years 1994 and 2001. Nineteen districts had adopted the D e p a r t m e n t of Health's target of 300 C A B G per million. 5 Twenty one of the responding districts had targets for PTCA. These ranged from 30 per million by 1995 to 650 per million with no target year specified. A few districts had also set target levels for coronary angiography and these were generally in the region of two angiograms for every P T C A or CABG.

Relationship between CHD SMR's and rate of CABG, P T C A and angiography No relationship was found between standardised m o r tality rates for C H D a m o n g s t those aged < 6 5 and crude rates of C A B G and P T C A combined (Spearman's rank correlation coefficient - 0 . 2 0 ; 95% CI - 0 . 4 8 to 0.11) and c o r o n a r y angiography (Spearman's rank correlation coefficient 0.06; 95% CI - 0 . 2 9 to 0.40) (Figures 1 and 2). Similarly no relationship was found with standardised mortality rates for those aged 65-74 (r = - 0 . 2 0 and r = 0.17 respectively). A weak relationship was found between the n u m b e r of C A B G ' s and the n u m b e r of P T C A ' s contracted for (Spearman's rank correlation coefficient 0.52; 95% CI 0.26 to 0.71 (Figure 3).

Table I Rates of coronary artery bypass grafting, percutaneous transtuminal coronary angioplasty and coronary angiography (per million total population) specified in District Health Authority contracts for 1993/4 Procedure

Mean rate per million

Median rate per million

Range

CABG PTCA Coronary Angiography Combined rate for CABG and PTCA*

374 183 1041

355 156 1010

162 to 710 18 to 648 581 to 2334

541

505

211 to 1258

* This value differs from the sum of the values for PTCA and CABG as some districts did not report PTCA rates: the reported value is for those districts reporting rates of both PTCA and CABG.

Invasive management CDH-purchasing I)J 8unnell and I Harvey 1200

Discussion

1000 o

800

~ E

...."..t'. ;.,..

e00

.%.

400 200 0

1

k

L

I

20

40

60

80

100

Standardised mortality rate per million for CHD in rr~n and wo~ aged<65

Figure 1 Relationship between standardised mortality rate per million for CHD (age <65; 1992) and crude rate of invasive treatment (CABG and PTCA; 1993/4)

Service developments Thirty eight (36%) of the responding health authorities r e p o r t e d having local angiography services available within their district (some at more than one centre). The mean number of coronary angiograms performed within each hospital per year was 1,085 (median 814; range: 120-5,562). F o u r units undertook fewer than 300 coronary angiograms per year. Eighteen (47%) of these services were not on the same site as cardiac surgery units. Hospitals within a further five districts were planning to develop coronary angiography services in the near future. Twenty-three health authorities reported a locally provided P T C A service; three of these had no on-site cardiac surgery back-up. A further four hospitals were planning to develop P T C A services, three of these without on-site cardiac surgery back-up. The mean number of PTCAs performed per year by each hospital was 724 (median: 652; range: 250-2,181). Twenty-one authorities had a provider of cardiac surgery within their district. Three further trusts planned to develop this service in the future. 2500

~

2000

1500

c

£

°

o 30(30 o o

500

i

0

10

20

30

40

50

60

70

80

90

S t l t n d a r d l ~ l mortal|~ rate per n~lllton

Figure 2 Relationship between standardised mortality rate per million for CHD (age <65; 1992) and crude rate per million population for coronary angiography (1993/4) 700 •

®

50o

e~ 400

**

300

• 0

100

200

3[}0

,o

* 400

500

600

700

CABG rate per million

Figure 3 Relationship between CABG rate per million and PTCA rate per million for District Health Authorities of England (t993/4)

This analysis was based upon the responses given by Directors of Public Health to a postal questionnaire. Responding districts were representative of all districts in terms of their standardised mortality rate for C H D and the responding D s P H were responsible for approximately t w o thirds of the population of the UK. Detailed information on contracting for 1993/4 was available for districts covering 47% of the population of the UK. Only two districts were unable to specify the number of procedures purchased by G P fundholders. Information on operations performed in the private sector was not obtained; such information would be unlikely to affect the main findings reported here although it is recognised that in some regions up to 20% of CABGs and PTCA's are performed outside the NHS. 4 Districts did not specify age and sex specific intervention rates in their contracts for cardiac services and therefore the data collected could not be used to calculate age and sex standardiscd rates of intervention. The crude rates of CABG, P T C A and angiography presented do however allow analysis of decisions regarding the balance of the purchase of CABG, P T C A and angiography. In addition a crude assessment of the relationship between mortality rates from C H D and purchase of C A B G and P T C A was possible, although age and sex standardised comparisons would have been preferable. The use of standardised mortality rates for C H D as a proxy measure of need for invasive treatment for C H D is not ideal. It is known, for example, that in men only 37% of incident c o r o n a r y heart disease is preceded by a history of angina, a Thus mortality from C H D is not necessarily related to the prevalence of angina in a community. Lastly some inaccuracies may have occurred in the responses to the questionnaire. One district, for example, reported undertaking 1 angiogram for every P T C A or C A B G it purchased implying an implausibly high predictive value for angiography. We have no reason to suspect systematic bias in the responses to our questionnaire. Despite these limitations the data support a number of conclusions. Firstly, there is marked variation between districts in levels of purchasing of CABG, P T C A and angiography. There was a four-f01d variation in the purchase of C A B G and angiography and a 36-fold variation in the purchase of PTCA. This is similar to the variation reported by the Clinical Standards Advisory G r o u p in four regions of the UK. The variation in rates for P T C A and C A B G do not occur as the result of preference for one procedure over the other; districts purchasing more CABG's also tended to purchase a greater quantity of PTCA's. There was no evidence that those districts where there was local provision of a CABG, P T C A or angiography service contracted for higher rates of investigation or treatment. Secondly the marked variation in rates of CABG, P T C A and angiography are mirrored by variations in target levels for these procedures in the different districts of the UK. M a n y districts have adopted with D e p a r t m e n t of Health's national target of 300 per million. N o national target has been agreed for P T C A and uncertainty surrounding the role of P T C A in the management of C H D was reflected in the 20-fold variation in target r a t e s for this procedure. This suggests

15

16

lnvasive management CDH,purchasing OJ Gunnelland I Harvey

that, for this procedure, the variations are likely to continue for some time in the future. Thirdly there was no significant relationship b e t w e e n rates of angiography, C A B G and P T C A and standardised mortality rates for C H D . Similar work in Northern Ireland has revealed no relationship between C H D mortality and angiography rates in 26 district council areas. 9 This contrasts with the weak inverse relationship between invasive procedures and C H D found by the Clinical Standards Advisory G r o u p (CSAG) in four regions of Britain (r = -0.35). Reasons for these differences are unclear. Lastly, variations were seen in the ratio of the number o f angiograms to invasive treatments purchased, although overall the ratio of invasive treatments to coronary angiography (i : 2) is that r e c o m m e n d e d by the British Cardiac Society. 1° The observed variations m a y be as a result of the increasing use of coronary angiography as a diagnostic test, variations in the appropriateness of the investigation and different approaches to the investigation of angina. 1~ Variations in the use of PTCA, C A B G and angiography suggests uncertainty concerning the appropriate indications for their use. This uncertainty is most marked in the case of PTCA, an intervention whose role in the m a n a g e m e n t of C H D is u n c e r t a i n ) Further evidence of this uncertainty comes from audits of the appropriateness the use of these interventions. An audit in Trent R H A in 1990 showed that 21% of angiographies and 16% C A B G s were rated inappropriate in relation to the patients clinical condition. ~2 Similar recent studies using American based consensus criteria for: appropriateness of CABG, P T C A and angiography in New Y o r k state have been reported. It appears that P T C A in p a r t i c u l a r is more often undertaken for inappropriate or uncertain indications, t 3 - t s Use of angiography is increasing, particularly with the development of district based services. Over a third of the responding districts now have local provision of coronary angiography, some with low rates of patient throughput. In one in nine districts without such a service local providers were planning its development. Although this figure m a y over-estimate the exteni of developments since those districts responding to our questionnaire m a y be those which have more recently reviewed their provision of cardiac services. It is likely that such developments will lead to increased inappropriate use of this intervention unless clear guidelines are developed. The rapid growth of district based angiography services should therefore be reviewed. The British Cardiac Society suggests that an important criterion for the development of such services is that they 'should provide services f o r more than one district to ensure that there is sufficient throughput to maintain local skills and to ensure cost efficiency'. 1°

This survey highlights a n u m b e r of problems faced by those purchasing treatments for C H D . The marked inter district variations in C H D mortality are likely to reflect differences in the prevalence of angina. However there is little information on the population incidence of angina and in particular on the incidence of those degrees of angina for which invasive m a n a g e m e n t is appropriate. This lack of information is exacerbated by uncertainty over both the a p p r o p r i a t e indications for C A B G and P T C A and the reIative long term cost effectiveness of these procedures. The preferred outcome for

angina patients too needs to be m o r e clearly defined; cost-effectiveness will vary depending u p o n whether the outcome is relief of angina, return to work, improved quality of life or reduced mortality. Faced with these uncertainties and the decreasing mortality rate from C H D in younger people purchasers m u s t balance levels of C A B G and P T C A against those for other treatments of proven benefit to patients. The observed variation reflects the outcome of a mixture of these decisions, historical precedent and supply led demand. Until further evidence becomes available purchasers should work with providers to develop guidelines on the m a n a g e m e n t of patents with C H D to ensure o p t i m u m use is m a d e of resources. Before the place of P T C A in the m a n a g e m e n t of C H D is m o r e clearly defined the rapid development of this treatment should be limited and target levels restricted. Acknowledgements

The authors would like to acknowledge the help of all the Directors of Public Health who took the time to respond to the questionnaire. References

1 Frankel S. The epidemiology of indications. Journal of Epidemiology in Community Health 1991 ; 45: 257-259. 2 Gunnell DJ, Harvey I, Smith L. The invasive management of angina: issues for Consumers and Commissioners. Journal of Epidemiology in Community Health 1995; 49: 335-343. 3 BARI, CABRI, EAST, GABI and RITA. Coronary angioptasty on trial. Lancet 1990; 335: 1315--1316. 4 Clinical Standards Advisory Group. Access to and availability of coronary artery bypass grafting and coronary angioplasty. HMSO: London, 1993.

5 King's Fund Concensus Development Conference Panel. Coronary artery bypass grafting. BMJ, 1984; 289: 15271529. 6 Fourth report of a Joint Cardiology Committee of the Royal College of Physicians of London and the Royal college of Surgeons of England. Provision of services for the diagnosis and treatment of heart disease. British Heart Journal 1992; 67: 106-116. 7 Department of Health. Public: Health Common Data Set 1993. HMSO: London, 1993. 8 Kannel WB, Feinleib M. Natural history of angina pectoris in the Framingham study: prognosis and survival. American Journal of Cardiology 1972; 29: 154-163. 9 Kee F. Referrals for coronary angiography in a high risk population. Quali~y in Health Care 1993; 2: 87-90. 10 A report of a working group of the BCS: Cardiology in the District General Hospital. British Heart Journal 1994; 72: 303-308. 11 Gray D, Hampton J, Variations in the use of coronary angiography in three cities in the Trent Region. British Heart Journal 1994; 71: 474--478. 12 Gray D, Hampton JR, Bernstein SJ, Kosecoff J, Brook RH. Audit of coronary angiography and bypass surgery. Lancet 1990; 335: 1317-1320. 13 Leape LL, Hilborne LH, Park RE, et al. The appropriateness of use of coronary artery bypass graft surgery in New York State..lAMA 1993; 269: 753-760. 14. Hitborne LH, Leape LL, Bernstein SJ, et at. The appropriateness of use of percutaneous transluminal coronary angioplasty in New York State, J A M A 1993; 269: 761-765~ 15 Bernstein SJ, Hilborne LH, Leape LL, et aI. The appropriateness to use of coronary angiography in New York State. J A M A 1993; 269: 766-769: