Rapid Responses to:
|
|
Rapid Responses: Rapid Responses published:
-
![[Read Rapid Response]](/icons/misc/sectionDOWN.gif)
Pragmatic screening
- Alexander Williams (7 September 2001)
-
Population screening for Diabetes IS cost effective
- Nigel Higson (9 September 2001)
-
Screening for diabetes in developing countries.
- V M Verma (9 September 2001)
-
Screening yield may not be quite so low.
- Mary Walker (12 September 2001)
-
Opportunistic screening for diabetes in general practice better than nothing at all
- Bertil Hagström (17 October 2001)
-
Cheap Diabetic Screening
- Philip Ramage (18 October 2001)
-
Diabetes screening in general practice
- Mel Bates (10 November 2001)
|
|
|||
|
Alexander Williams, General Practitioner St Thomas Health Centre
Send response to journal:
|
Dear editor, Lawrence and colleagues(1)suggest that screening for diabetes in patients over 45 has a low yield of only 0.2%, only 876 of 2481 patients (35%) attended for screening,in their study. All patients can be offered a new patient health check when they register at a new surgery for which a fee of £8.20 is paid. A new principle in general practice is able to offer this service to all patients on their list.When i became a new principle in 1982 i tried to offer this health check to all patients as they attended the surgery as it allowed me to update their medical records and perform some simple baseline screening.I am not sure how many patients i was able to screen in 6 months but identified 5 new cases of diabetes on urine tasting for glycosuria,and this equates to a prevalence of about 0.25% on a list size of 2000 and compares to the rates that lawrence quotes. Given the low attendance rates for screening in lawrences study and a financial inducement to offer a new patient health check this may be a more pragmatic approach to population screeing reserving fasting plasma glucose for those who have glycosuria or are at high risk. 1.Lawrence JM,Bennett.P,Young A, RobinsonAM. Screening for diabetes in general practice:cross sectional population study. BMJ 2001;323 :548-51 |
|||
|
|
|||
|
Nigel Higson, GP Goodwood Court Medical Centre, 52 Cromwell Road, Hove BN3 3DX
Send response to journal:
|
I was pleased to read this article by Lawrence et al - it raises two important aspects of primary care - the first is that I would question the ability to conduct a consultation in 10 minutes which involved full and informed consent and discussion of implications of diagnosis, together with body mass measurement, full family history and phlebotomy... I would suggest 20 minutes would be the minimum required hence doubling consultation costs. It may well be that in the West Country practitioners talk at a higher speed. The second point is the suggestion that the process might be cost effective if selective screening was offered. Primary care records are not yet sufficiently standardised or historically coded to the extent that selective screening could occur. We have, over the last few years repeatedly offered patients home self-screening followed by fasting blood- sugar estimation for those demonstrating glycosuria two hours after a glucose load. This has repeatedly discovered previously unknown diabetic patients at a minimal cost per new diabetic. Self-screening by random urine tests was not felt to be worthwhile since it would not identify those people with impaired glucose tolerance. By asking the patient to consume a high glucose drink, followed by a self- administered urine test two hours later, we believed a number of unknown diabetic patients would be identified. In April 2001, 4,000 newsletters were sent out by second-class mail to all adult patients registered at Goodwood Court Surgery - one per household. Of these, 2,000 included urine testing strips. Two months later the surgery had been contacted by 27 patients reporting a positive urine test result who were then invited to have a fasting (14 hours) blood sugar estimation. The samples from these further tests were sent to the local hospital laboratory. @bt:Six patients were found to be diabetic (blood sugars 7.9;7.2;7.9;9.0;9.8;10.5mmol/l) and one patient had blood sugars within 0.5mmol above the upper range of normal for the laboratory(6.3mmol/l). The non-diabetic patient with high blood sugar will be recalled on a 6-monthly basis for repeat fasting blood sugars. Urine testing strips are relatively cheap - 2,000 were purchased for £120. Time spent in placing sticks in the envelopes, sealing and labelling took five hours (£30). Envelopes and labels cost £25. Twenty seven laboratory estimations were required as well as nursing time in phlebotomy (£270) Taken together the cost of screening 2,000 patients for diabetes with a diagnostic success rate of 0.3% cost £805 including postage = £134 for each new diabetic identified. This would decrease to £70 if the cost of postage was excluded (in our case, letters were being sent to patients for other reasons and this was not an additional expense) Many patients have yet to be identified as being diabetic (hyperglycaemic) within the UK. Screening by blood analysis may deter some patients and is expensive in both time and money. Urine testing is cheap and can be undertaken by the patient or carer with no formal training. Prior loading with a glucose drink is more likely to demonstrate inability to handle glucose with glycosuria. Commercially available glucose drinks are available and are acceptable to patients. An effective self-testing programme based on glucose loading and urine testing could save the NHS some long-term treatment costs and decrease secondary morbidity associated with age-related diabetes. This study has shown repeatedly over a five year programme that it is able to identify previously unknown hyperglyaemic patients at low cost with sensitivity down to fasting blood sugar levels of 7.2mmol/l. |
|||
|
|
|||
|
V M Verma, Medical Practitioner, Delhi, India India at present
Send response to journal:
|
I agree with Dr. Alexander's approach of using glycosuria, especially for routine screening for diabetes in macro-level populations in developing countries.Screening for diabetes for medical insurance in India, is always done by an initial urine for sugar test...more economical, and user-friendly rather than using "pricking" methods. regards, Mona DrVM Verma |
|||
|
|
|||
|
Mary Walker, Senior lecturer in Epidemiology Department of Primary Care and Population Sciences, Royal Free and UCLMS, London NW3 2PF
Send response to journal:
|
Editor, I was concerned to note that in the paper from Lawrence et al Sept 9th 2001, their screened subjects had a mean BMI of 25.4: SD 4.1 and unscreened subjects a mean BMI of 26.1: SD 4.4. This difference is highly significant p<0.001, but they do not mention this. Indeed, their unscreened subjects are likely to contain a higher % of undiagnosed diabetics on this factor alone. In addition they appear to have changed the criteria of the plasma glucose categories in their table 2 ( 6.1- <7.1mmol/L, and =>7.1mmol/L)from those recommended by WHO (6.1mmol/L -<7.0mmol/L for IFG and =>7.0mmol/L for diagnosing diabetes). |
|||
|
|
|||
|
Bertil Hagström, lecturer, GP Department of Primary Health Care, Göteborg University, Sweden
Send response to journal:
|
Editor In the paper by Lawrence et al (1) the authors conclude that screening for diabetes in general practice by measuring fasting blod glucose has a very low yield in patients whose sole risk factor for diabetes is age 45 or more. Our experience from opportunistic screening in a rural primary health centre (PHC) in Sweden is different. By a sign in the waiting room during 1999 all visitors 40 years old or more at PHC in Storvik, a village of 6800 inhabitants, were offered to test their blood glucose. 249 patients accepted, of whom 72 had a non fasting capillary blood glucose (non-FBG) of >6,7 mmol/L. They were recommended to come back for two FBG tests. Sixtytwo persons showed up and 18 of them had a FBG of > 6,1 mmol/L in both tests (2,3). Five persons were around 50 years old and eight were 75 years or older. 349 tests were performed alltogether. The expenses for material and work per test were about two £ (4). Thus, by an opportunistic screening we found in our study to very low costs 18 new diabetic patients. Our opportunistic way of screening to detect early diabetes was simple. £ 40 to find a diabetic person or about 140£ per person to find these five ”younger” persons are low costs. To set the cut-off level for a normal random blood glucose to < 6,7 mmol/L is said to give a sensitivity of 64% and a specificity of 92% in a well screened population (5). Since most people visit the PHC in a five years time, one will have reasonably few missed cases. Therefor this is an acceptable way of screening and without any oral glucose tolerance tests. We suggest it is better to do something than nothing at all and we offer all patients a blood glucose test at least every five years from the age of 45. Patients with cardiovascular risk factors, such as hypertension and dyslipidaemia, are tested at regular visits at least yearly. References 1. Lawrence J M, Bennett P, Young A, Robinson A M. Screening for diabetes in general practice: cross sectional population study. BMJ 2001;323:548-551. 2. Report of the Expert Committe on the diagnosis and classification of diabetes mellitus. Diabetes Care 1997; 20 1183-97. 3. Mayfield J. Diagnosis and classification of diabetes mellitus: new criteria. Am Fam Physician 1998. 58; 6: 1355-62, 1369-70. 4. CDC Diabetes Cost-Effectiveness Study Group, Centers for Disease Control and Prevention. The cost-effectiveness of screening for type 2 diabetes. JAMA 1998. 280; 20: 1757-63. 5. Andersson D, Pettersson C. Screening in primary health care. How do the new criteria affect the incidence of diabetes? Lakartidningen 1998. 95; 46: 5171-5. Swedish. Bertil Hagström, Bengt Mattsson, |
|||
|
|
|||
|
Philip Ramage, General Practitioner (defeated) Warlingham, Surrey CR6 9DE
Send response to journal:
|
Sir, The issue of the practicalities of screening for diabetes are with us again. May I resurrect an idea I put years ago to both the DoH and the BDA (ignored by both). I suggest that with every Sunday newspaper is enclosed a urine testing strip for glucose. These, such as have been made available in the past for GPs, can be single-packaged, and could include simple instructions. I suspect that one or many of our news conglomerates would be happy to be associated with such a philanthropic venture, and the costs could be minimal. We might achieve national awareness of and access to a household-wide pool of fresh "stix", the elderly could be tested regularly and the young - notoriously reluctant to attend their GP - could reassure themselves on any day they wondered about their diagnosis. As they say - this isn't rocket-science. Whatever the epidemiological niceties of screening, glycosuria requires investigation under any circumstances, and identifying it this way would cost nothing. |
|||
|
|
|||
|
Mel Bates, General Practitioner Fairview Family Practice,
Send response to journal:
|
Dear Editor, As the risks of diabetes are coming to be more keenly appreciated and as the benefits of good glycaemic control are now established beyond reasonable doubt, several authorities, including the Irish College of General Practitioners are advocating that GPs should be more actively seeking out patients with undiagnosed diabetes 1-3. Lawrence et al have highlighted the potential workload implications of implementing guidelines regarding diabetes screening in general practice 4. We have also carried out recent research, which highlights the consequences of a more proactive approach to diabetes screening in one Dublin practice. We screened all those 65 and over who presented for their "flu" vaccine in the winter of 2000 using Glucotrend devices. They were validated beforehand and once weekly during the project. Of the 339 who attended for influenza vaccination (mean age 76), 29 (8.5%) were known diabetics and 3 had previously been found to have impaired fasting glucose or glucose tolerance. 2 patients declined and 2 were excluded from the study on the grounds of dementia. Therefore, 303 patients were eventually screened for diabetes. Of these, 64 (21%) patients had a glucometer reading >or equal to 6mmol/l. These were all offered an abbreviated glucose tolerance test (GTT) but in 18 cases this was not done (7 declined, 2 deceased , 1 investigated in hospital , 9 nursing home/housebound ) . Of the 7 who declined, 5 did allow a fasting blood glucose of which one was found to be impaired. Of the 46 glucose tolerance tests performed, 28 were normal, 6 revealed frank diabetes and 12 revealed previously undetected impaired glucose tolerance. 103 (34%) patients screened had a glucometer reading of > or equal to 5.5mmol/l which is, actually, the recommended cut-off point in the current Irish guidelines, based on Irish and Australian diabetes screening studies 5 6. Allowing 5 minutes for initial glucometer reading including explanations (303x5=25 hours) and 8 minutes for FBG (8x5=40 mins) and 20 minutes per GTT (46x20=15hours), screening 303 patients using 6mmol/l cut-off point involved over 40 hours of additional work . Taking 5.5mmol/l as our cut-off would have added 13 hours (39x20). This extra work identified 6 new diabetics and 13 people with impaired fasting glucose or impaired glucose tolerance but this yield is on a background of a fairly high proportion of our diabetics already being known. Several studies and our local guidelines suggest a cut-off point for random glucose testing of 5.5mmol/l. Even excluding our housebound patients and using 6 mmol/l as our cut-off point resulted in a considerable extra workload. We strongly suggest that the issuing of guidelines should be preceded by workload studies as well as being based on clinical trial evidence so that resources can be put in place before their implementation to enable GPs to adhere to them. Yours sincerely, M Bates1, P Carmody1, S Haba2, S Smith2 and C Bradley3. 1Fairview Family Practice, Dublin 3 2Department of General Practice, University College Dublin 3Department of General Practice, University College Cork Address for correspondence: Dr Mel Bates, Fairview Family Practice, 17 Fairview Strand, Dublin 3, Ireland Email: batesm@gofree.indigo.ie References 1. Alberti K, Gries F, Jervell J, Krans H. A desktop guide for the management of Non-insulin dependent diabetes mellitus (NIDDM): an update. Diabetic Medicine 1994;1994(11):899-909. 2. Engelgau MM, Narayan KM, Herman WH. Screening for type 2 diabetes [In Process Citation]. Diabetes Care 2000;23(10):1563-80. 3. O'Sullivan T, HArkins V, Houlihan J. Guidelines for diabetes care in the community. Dublin: Irish College of General Practitioners, 2000. 4. Lawrence JM, Bennett P, Young A, Robinson AM. Screening for diabetes in general practice: a cross sectional population study. British Medical Journal 2001;323:548-551. 5. Firth R, Smith S, Toal J, Doherty P, Hamalian F, McAuliffe A, et al. The Irish Diabetes Detection Programme in General Practice. Diabetologia 2000;43(S1):I-IV. 6. Welborn TA. Diagnosing Australians at risk of diabetes mellitus in general practice. Med J Aust 2000;173(2):61-2. |
|||