Family Practice Advance Access originally published online on April 1, 2005
Family Practice 2005 22(3):317-322; doi:10.1093/fampra/cmi014
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The costs of care in general practice: patients compared by the council tax valuation band of their home address
a Northlands R & D General Practice, Calne, Wiltshire, b Division of Primary Health Care, University of Bristol and c Bath & Swindon RDSU, Royal United Hospital, Bath.
Correspondence to Dr Norman Beale, Northlands Surgery, North Street, Calne, Wiltshire SN11 OHH, UK; Email: info{at}northlandsrnd.org.uk
Received 26 April 2004; Accepted 22 December 2004.
Beale N, Hollinghurst S, Taylor G, Gwynne M, Peart C and Straker-Cook D. The costs of care in general practice: patients compared by the council tax valuation band of their home address. Family Practice 2005; 22: 317–322.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodResultsDiscussionDeclarationReferences |
|---|
Background. It is difficult to measure and compare workload in UK general practice. A GP/health economist team recently proposed a means of calculating the unit cost of a GP consulting. It is therefore now possible to extrapolate to the costs of other clinical tasks in a practice and then to compare the workloads of caring for different patients and compare between practices.
Objectives. The study aims were: (i) to estimate the relative costs of daily clinical activities within a practice (implying workload); and (ii) to compare the costs of caring for different types of patients categorized by gender, by age, and by socio-economic status as marked by the Council Tax Valuation Band (CTVB) of home address.
Methods. The study design was a cross-sectional cost comparison of all clinical activity aggregated, by patient, over one year in an English semi-rural general practice. The subjects were 3339 practice patients, randomly selected. The main outcome measures were costs per clinical domain and overall costs per patient per year; both then compared by gender, age group and by CTVB.
Results. CTVB is as significant a predictor of patient care cost (workload) as is patient gender and age (both already known).
Conclusions. It is now possible to estimate the cost of care of different patients in such a way that NHS planning and especially resource allocation to practices could be improved.
Keywords. General practice, workload comparisons, costs of care, council tax bands.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionDeclarationReferences |
|---|
In a recently reported study we examined the validity of Council Tax Valuation Band (CTVB) of patient residence as a predictor of face to face contact rates with our patients.1 This showed a clear and significant relationship between CTVB and such consultations, the highest rates being for CTVB ‘A’ patients with a progressive drop to those in bands ‘E’ and above, a correlation not confounded by gender or age. We recognise, however, that consultations represent only a part of the complex clinical ‘work’ that engages, daily and ceaselessly, every UK general practice. Patients will ring the surgery for advice, request repeat prescriptions, require referral to secondary services, attend for clinical investigations; they can make unpredictable demands at anti-social times. In fact the complexity of activities is difficult to classify and comprehend, even in a book.2 Even more problematic is the application, to all of these activities, of an acceptable common unit of workload that would permit cross-comparison. In terms of the time and effort involved, the formulating and issuing of 20 repeat prescriptions might be equivalent to a single consultation with a doctor. The writing and transmission of a request for an outpatient consultation might match the taking of two blood samples for the local laboratory. No one knows, and no one can know, unless one agrees a common currency of ‘work’. The only real possibility seems to lie in attributing a monetary value to each and every clinical task.
Thanks to the work of Netten and Curtis3 on the unit costs of health and social care and then of Kernick and Netten who have gone on to propose a framework to enable individual practices to determine their unit consultation costs,4 this is now feasible. It is also possible, then, to extend their technique to make a monetary assessment of other clinical domains. We can then calculate a putative ‘cost’ for each item of workload and, by aggregating these, the ‘cost’ for each patient in a specified time: we can then begin to assess and contrast the burdens engendered by different types of patient.
In 1992 the British Government replaced the Community Charge (‘Poll Tax’) with a new tax—the Council Tax.5 This was to be based on an external assessment of each house. Each was to be allotted an ‘open market’ value, as at 1 April 1991, based on size, layout, character and locality, and placed into one of eight ‘Valuation Bands’, A–H, that would dictate the amount of the tax. All UK Local Authorities were mandated to levy the new tax and to publish lists showing the Council Tax Valuation Band (CTVB) of all properties in their jurisdiction: these are now available, for England and Wales, on a web site published by the Valuation Office Agency.6 We first compared this new ‘ecological attribute’ of all patients, the CTVB of their residence, with a compound measure of morbidity in a small study reported in 2000.7 We demonstrated an association between CTVB and:
- the number of rooms per house;
- established deprivation indicators viz. home ownership, access to a car; and
- workload in a typical UK general practice, i.e. elements of recorded morbidity (doctor consultations, home visits, investigations and referrals).
- established deprivation indicators viz. home ownership, access to a car; and
In this study we aim to estimate, extrapolating from the model by Kernick and Netten,4 the relative costs of the many clinical activities that encompass the work of a general practice; and to compare, for one year, these costs of care aggregated for patients categorised by gender, age, and socio-economic status as marked by the CTVB of their home addresses.
|
Method |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionDeclarationReferences |
|---|
The practice
Calne is a market town that became industrialised. Reference to the Office for Government Statistics Website9 shows the spectrum of Council Tax Bands for the Calne neighbourhood to be skewed towards the lower bands respect the regional average but otherwise representative. The same official data show that there are no significant ethnic minorities.
Northlands is a semi-rural Research and Development/ Training practice in which six partners (two part-time) work from modern, purpose-built premises. List size is currently 10 300 (as at end of study period). Its complements of nurses and administrative staff are above local averages but it is, otherwise, a typical general practice: age spectrum and annual turnover are close to local and national averages.
Study design
Clinical entries in the medical records (FP5s & FP6s) of one third of the practice patients (selected alphabetically) were the study data source. Morbidity markers were extracted manually and transferred, coded, onto computerised files that had been pre-entered to record, for each patient:
- name, forename, gender (encrypted);
- date of birth;
- age at end of study period;
- doctor with whom registered;
- home address and post-code;
- enumeration district and UPA8 score;
- distance from home address to surgery (more or less than 3 miles);
- years registered with practice;
- months registered during study period.
- face to face consultations with a doctor (excepting d, f, g as below);
- face to face consultations with a practice nurse (excepting d, e, i, j);
- ‘DNA’s—those instances in which a patient had made an appointment with a doctor or a nurse but failed to attend without cancellation;
- ‘triage’ A consultations—always face-to-face with a nurse but initiating a further, immediate consultation with a doctor;
- ‘triage’ B consultations—always face to face with a nurse but resolved by the issuing of a prescription without doctor consultation, by recommending a later, routine, appointment with a doctor, or by giving appropriate other advice;
- ‘pool’ consultations with a doctor—those consultations organised, mostly at the end of the working day, for those patients demanding to be seen that day and for which no other option was available;
- home visits to patients in their own homes;
- telephone calls made, by patients, to the practice;
- nurse consultations held in designated chronic disease management clinics;
- clinical investigations performed by practice nurses—mostly venepuncture samples;
- referrals to secondary services, whether immediate or by appointment;
- acute drug prescriptions issued;
- repeat drug prescriptions issued;
- out-of-hours contacts made by patients—whether response was telephonic advice, seen on attendance at primary care centre or visited at home.
Cost analyses
The cost analysis was conducted from the viewpoint of the general practice. We estimated the average resource use for each of the fourteen clinical domains, in terms of time spent by a doctor, a practice nurse, and administrative support. Where necessary, a member of the research team conducted a stopwatch exercise to arrive at a mean length of time for each activity. We estimated cost per minute for GPs and practice nurses using the framework devised by Netten and Curtis,3 with adjustments to tailor the published estimates to fit this particular study, as proposed by Kernick and Netten.4 Cost per minute of administrative support was estimated from practice accounts. The unit costs were applied to resource use estimates to arrive at a cost per clinical domain, and these were used to value each patient contact to arrive at a total cost per patient. A full description of the cost analysis is given in Appendix 1.
Statistical testing
Costs were analysed per patient years at risk. Rates were then compared between groups defined by gender, age-band, length of registration with the practice, recent registration with the practice (within one year of the beginning of, or within the study period), distance of home address from the surgery (up to three miles versus more than three miles), UPA8 score, and CTVB. Variables were initially considered in a univariate manner using Spearman's rank correlation coefficient. Significant variables thus identified were then included in multivariate tests using an exponential regression model. All analyses were performed in SPSS version 12.0.
|
Results |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionDeclarationReferences |
|---|
After omitting 39 institutionalized patients (1.2%), the study group consisted of 3343 patients, 1647 female and 1696 male. It was not possible to allocate a CTVB to 4 patients (0.1%) and a UPA8 score to a further 179 patients (5.4%). Therefore data retrieval was achieved, at worst, for 3153 (94.6%) of the potential study group.
The mean annual costs per patient in conventional age bands are shown in Table 1. Costs are high for young children, fall in youth, and then increase progressively with age, as one would expect. The mean annual costs per patient aggregated by CTVB (all patients, males and females separately) are presented in Table 2: CTVBs E, F, G, H were aggregated into an ‘E+’ group to obviate small numbers. Costs for females are higher than for males as would be expected. There are also cost gradients across the CTVB spectra, more pronounced in the females. Mean annual costs for each clinical domain are shown (all patients) in Table 3 (where rows are sorted by mean cost, most expensive first).
|
|
|
In the univariate tests on rates for between-group variables, gender, age, recent practice registration, and CTVB were shown to be significant. ‘Years registered with the practice’ was also a significant variable. UPA8 was not significantly associated. The multivariate analysis confirmed significant associations (all at P < 0.001) for gender, age, CTVB, and recent registration.
The new finding is the significant association between annual costs of care for patients and the CTVB of their home addresses, costs decreasing progressively as CTVB increases.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionDeclarationReferences |
|---|
This study fulfils its broad objectives. It is a demonstration that the relative costs of the many aspects of modern UK general practice can be estimated and compared. The results have a high face validity and also show that different patients engender different overall costs. They support the suspicion that there is a factor other than gender, age, length of registration (which is probably a proxy for age itself) and list turnover that plays a significant part in governing workload. This is socio-economic status and although there is already evidence for this influence, the study demonstrates anew the significance of this factor, in women especially, by using CTVB as a proxy marker; a marker that is universally available, stable (indeed official) and can be applied at individual and household level.
While it is true that the study is a small one for this kind of research—epidemiological—the findings are consistent and statistically significant. We also acknowledge the limitations of the study base being in one general practice, in an extremely small fragment of the NHS. There may be very different costs in running general practices in different kinds of community and in practices of different sizes—all this remains to be explored. On the other hand there is no reason to consider the study setting to be other than a typical working practice, at least for a semi-rural environment where there is no ethnic minority and all novel ideas deserve at least one test-run on a small scale. Study method might also be criticised: we certainly recognise that costing of activities is very difficult and that assumptions have to be made. Nevertheless we have tried to be rigorous and, finally, we draw only comparative conclusions rather than promote the calculated costs themselves. It is also for this reason that we saw no relevance in performing sensitivity analyses.
We trust that the study may inspire further investigation—at multi-practice level. It certainly suggests further test hypotheses. For instance we see, perhaps for the first time, the relative effort required to match the daily demands in the now complex structure of everyday practice. Face to face consultation with patients is the core activity (Table 3), as would be expected, but there are many other relentless daily activities in any practice and their burdens relative to each other will need corroboration. There is also a strong hint that the socio-economic gradient demonstrated by CTVB is consistently present in those clinical domains initiated mostly by patients such as doctor consultations or out-of-hours contacts. This contrasts with the absence of any such consistent gradient for activities engendered by clinical staff, such as investigations or referrals to specialist care. Here, in embryonic form, might be the means of measuring differential professional responses to clinical demands. Do some types of patient expect and receive more attention and resource than their level of morbidity deserves—do some patients get a better deal from their doctors?
Finally, we see the power of, and the need for, a health economics perspective in considering what GPs do. There has been little work on the microeconomics of general practice although there is now a rapidly increasing interest in this area following the appearance of a new contract for general practice.10 This incorporates the concepts that GPs will want to be able to limit their workloads, be paid more rationally for what they do, and that health authorities will want evidence of increasing clinical quality. However, the contract was very nearly stillborn—for want of the very sort of fiscal and comparative information that this study provides.
|
Declaration |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionDeclarationReferences |
|---|
Funding: Dr Beale is supported by an NHS Research and Development Support Grant.
Ethical approval: granted by the Bath Local Ethics Committee (BA004, 2001).
Conflicts of interest: none.
|
|
Acknowledgments |
|---|
We thank everyone at Northlands Surgery for their cooperation, The Board of Science and Education of the BMA for financial support (Dr Beale was awarded the Joan Dawkins Award—Housing and Health for 2001), and the Bath and Swindon RDSU.
|
References |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionDeclarationReferences |
|---|
1 Beale N, Taylor G, Straker-Cook D, Gwynee M, Peart C. Council tax valuation band of patient residence and clinical contacts in a general practice. Br J Gen Pract 2005; 55: 32–36.[Medline]
2 Waller J, Hodgkin P. General practice: demanding work. Oxford: Radcliffe Medical Press; 2000.
3 Netten A, Curtis L. Unit Costs of Health & Social Care 2002. Personal Social Services Research Unit, University of Kent, Canterbury (2002).
4 Kernick D, Netten A. A methodological framework to derive the cost of the GP consultation. Fam Pract 2002; 19: 500–503.
5 Anonymous. Local Government Finance Act 1992. Council Tax. HMSO: London; 1992.
6 http://www.voa.gov.uk/council_tax/index.htm (last accessed 16/4/04).
7 Beale N, Baker N, Straker-Cook D. Council tax valuation band as marker of deprivation and of general practice workload. Public Health 2000; 114: 260–264.[Medline]
8 Beale N, Taylor G, Straker-Cook D. Is Council Tax Valuation Band a predictor of mortality? BMC Public Health 2002; 2: 17–24.[Medline]
9 http://www.statistics.gov.uk (last accessed 16/4/04).
10 Anonymous. New GMS Contract 2003. London: BMA; 2003.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  S. Hollinghurst, D. Sharp, K. Ballard, J. Barnett, A. Beattie, M. Evans, G. Lewith, K. Middleton, F. Oxford, F. Webley, et al. Randomised controlled trial of Alexander technique lessons, exercise, and massage (ATEAM) for chronic and recurrent back pain: economic evaluation BMJ, December 11, 2008; 337(dec11_2): a2656 - a2656. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Hollinghurst, N. Redmond, C. Costelloe, A. Montgomery, M. Fletcher, T. J Peters, and A. D Hay Paracetamol plus ibuprofen for the treatment of fever in children (PITCH): economic evaluation of a randomised controlled trial BMJ, September 9, 2008; 337(sep09_3): a1490 - a1490. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||