• 
• 

Three Part Question

Is [estimation] of [body weight] by [health care professional in the emergency department] accurate enough for correct drug dosing?

Clinical Scenario

A 65 year old gentleman presents with his wife after collapsing at home. He is FAST positive with a clear onset time of 1 hour. On examination in the emergency department he has a dense right hemiplegia and expressive dysphagia. CT shows no bleed. Stroke thrombolysis is considered however the patient is unable to tell you how much he weights and his wife is unsure. The doctor and nurses looking after him think he is around 70kg. Is this an accurate enough estimation for drug dosing?

Search Strategy

2 resources were selected for literature search, Ovid MEDLINE® 1948 to February Week 4 2013 and EMBASE® 1974 to 2013 February 27

The following search was performed:
([estimation.mp.] OR [estimate.mp.] OR [approximation.mp.] OR [assessment.mp.]) AND ([weight.mp.] AND [body.mp.]) AND ([doctor.mp.] OR [physician.mp.] OR [nurse.mp.] OR [nursing staff.mp.] OR [medical staff.mp.] OR [emergency department.mp.] The searches were limited to English language.

Search Outcome

In total 390 article were found form the OVID Medline search, 17 articles were found to be relevant. A further 1148 articles were found from the EMBASE search. In this search a further 2 relevant articles were found. In total 19 relevant papers have been assessed.

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Dos Reis Macedo et al July 2011 Brazil	28 patients in need of parental anticoagulation had their body weight subjectively estimated in the emergency department. Basal and steady-state (after the second subcutaneous shot of enoxaparin) anti-Xa activity was obtained as a measure of adequate anticoagulation.28 patients in need of parental anticoagulation has their body weight subjectively estimated in the emergency department. Basal and steady-state (after the second subcutaneous shot of enoxaparin) anti-Xa activity was obtained as a measure of adequate anticoagulation.	Prospective Observational study	Correct dose of more than 0.9mg/kg	From the 28 patients enrolled, 75% (group 1, n = 21) received at least 0.9 mg/kg per dose twice daily and 25% (group 2, n = 7) received less than 0.9 mg/kg per dose twice daily of enoxaparin.	Small number of patient. Unequal groups. Didn’t measure clinical events. Rounding up of dose due to the commercial available presentation of drug
			Anti Xa activity	Only 4 (14.3%) of all patients had anti-Xa activity less than the inferior limit of the therapeutic range (<0.5 UI/mL), all of them from group 2.
Brueur L et al Dec 2010 Germany	109 patient receiving intravenous thrombolysis patients were prospectively included in the study (only 100 patients had complete data sets). Body weight was estimated independently by 2 physicians, 2 emergency nurses, and a neuroradiological technical assistant. Patients were weighed, and anthropometric measurements for body weight approximation were taken. Dosing errors were assessed. Clinical outcome was evaluated at 90 days.	Prospective Observational Study	Estimation > 10% under or over actual weight	There was a greater than 10% discrepancy in 20.8% of patients estimates, 38.2% of the treating physician’s estimates, 25.9% of relatives estimates, 44.4% of neuroradiological technicians estimates, 42.2% of the emergency nurse’s estimates and 20% of the anthropometric measurements estimate.	Small number of observers. Patients weight 24 hours following treatment. Low case number. Only one centre therefore may not be generalisable.
			2 Observers	Mean of 2 observers improved estimates. 2 doctors estimates used then estimate improved from 38.2 to 34.6 and mean of 2 nurses improved estimate from 42.2 to 33.3%.
			Incorrect dosage	29% of patients were given a dose 10% more or less than the optimal dose
			Dosage error Effect	Under-dosage was an independent predictor for worse outcome in multivariate analysis. However over-dosage didn’t lead to an increased risk of bleeding.
Maskin LP et al Sep 2010 Argentina	42 consecutive patients were weighed by a physician with a calibrated stretcher scale. Attending physicians and nurses blinded to measured weight were asked to estimate patient\\\\\\\'s actual weight and height. Using the estimated height they predicted the body weight was calculated using the ARDSnet formulae (formula using height to estimate weight).	Observational study	Visual assessment	Overall accuracy within 10% error was 56%. Under estimated by greater then 20% in 15.6% of cases.	Repeat estimations by practioners of different patients may have resulted to increased attention. Small numbers of observers.
			Practioner	Nurses more accurate then physicians 63% vs 53%. Not statistically relevant.
			Predicted weight using ARDSnet	Overall accuracy of predicted weight using estimated heights to within 10% error was 40%. Predicted weight were generally overestimated.
Lin BW et al Nov 2009 USA	Prospective study of 235 adult ED patients. Mid-arm circumference and knee height were measured. These values were input in to equations to calculate patients\\\' weights. A physician and nurse were then independently asked to estimate the patients\\\' weights. Each patient was asked to report his/her own weight before being weighed.	Cohort observational study	Accuracy within 10% of actual weight	Calculated estimates were 69% accurate, Physician estimates were 54% accurate, nurse estimates were 51% accurate and patient estimates were 86% accurate.	Not all patients removed clothes, only two equations i.e. male and female. Race and age not accounted for. Data collected between 9 - 5
			Male and female	Weight estimation tool calculated weights more accurately in males (74%), than females (65%).
Spicer K et al Mar 2009 UK	As part of this study 50 healthcare staff (8 doctors, 34 nurses, 8 medical students) made 533 weight estimations on 182 patients. These estimates took place in three locations: the coronary care unit, cardiac catheterisation laboratory, and the cardiac outpatient department.	Cohort observational study	Patient vs Healthcare staff	Patients were more accurate than healthcare staff at estimating their weight to within 5% (80% vs 39%)	Bed bound or morbidly obese excluded as unable to weight on scales. Results reflect practice in a single institution and therefore may not be generalisable.
			Weight	There was a tendency to overestimate the weight (>5%) of lighter patients (BMI <25) and underestimate (<5%) the weight of heavier patients (BMI > 30)
			Sex of patient	Female patients were more likely to be accurate than men (62% vs 44%, p=0.035).
Lorenz MW et al Dec 2007 Germany	Formulae for estimating body weight from anthropometric measurement constructed using general population sample of nearly 7000 subjects. Use of formulae compared with the best visual estimation of two experienced physicians in a sample of 178 consecutive inpatients admitted to stroke unit.	Cohort study	Greater then 10% mismatch.	Best weight estimation from the 2 physician greater than 10% inaccurate in 31.5% and 40.4% patients. The formula was >10% inaccurate in 6.2% and 7.3% of patients depending on the formula used and the patients estimate was >10% inaccurate 3.4% of patients	Formula constructed on measurments taken from standing patients. For validation study patients were assessed supine. Different physicians used to give the 2 estimates
			Multiple physician	The combined best estimation of two physician were >10% inaccurate in 15.2% of patients.
Kahn CA et al Mar 2007 USA	Eighty-seven staff provided 957 estimates on 241 patients. Patients were weighed at the bedside, then staff were asked individually for estimates.	Observational cohort study	Estimation to within 10%	Providers were within 10% of measured weight on 63% of estimates	Estimates unwitnessed. 2 analog scales used. BMI calculated on self reported height
			Effect of weight of patient on estimate	A significant difference was noted in estimation of patient at different body mass index (BMI); percentages of correct estimates were 16% for BMI < 18.5; 38% for 18.5 < or = BMI < or = 30 and 23% for BMI > 30. BMI was the only statistically significant factor contributing to error.
Bloomfield R. et al 2006 Aug UK	Prospective study of 14 patients who had their height and weight estimated by 20 members of the medical and nursing staff (7 senior medical, 7 junior medical and 6 nurses). After all estimates had been recorded, measurements of weight and height were obtained. Weight was measured by means of a patient hoist with a calibrated weighing facility and height using a steel tape measure.	Prospective clinical study	Weight Estimation	Estimation of weight was poor, with 47% of estimates at least 10% different and 19% of estimates at least 20% different from the measured values. Estimation of weight was poor, with 47% of estimates at least 10% different and 19% of estimates at least 20% different from the measured values.	Small number of patients
			Weight of Patient	There was a trend to overestimate the weight of patients of low body weight and underestimate individuals at the other end of the weight range.
			Grade	No significant different compared with grade of doctor/nurse.
Menon S et al Apr 2005 Australia	Study of 1137 patients presenting to ED. Medical staff, nursing staff and patients were asked to estimate patient weight that was then measured.	Observational study	% error in estimates	Average per cent error in estimates was 3.9% for patients, 7.7% for nurses and 11% for physicians.	Patients not always weighed before health care worker estimates. No comment on ethnicity.
			10% of actual weight	91% of patients, 78% of nurses and 59% of physicians made weight estimates accurate to within 10% of actual weight.
Corbo J et al Mar 2005 USA	458 patients presenting to the emergency department estimated their own weight and had their weights estimated by treating physician and nurse	Prospective observational study	Weight estimates	91% of patient weight estimates were within 10% of actual compared to only 50% of physician and nurses estimates.	Patients unable to stand and those unable to estimate own weight were excluded. Patient recruitment limited by availability of research team.
			Weight of patient	Lighter patient were overestimated while heavier patient were underestimated.
			Experience of physician	Experience of physician did not appear to increase accuracy.
Angle-meyer BL, et al Nov 2004 USA	Prospective study enrolled a sample of 394 medically stable adults. The patient’s attending physician, resident physician, nurse, a paramedic, and the patient estimated the patient’s weight.	Cohort observational study	Estimation > 20%	Patients erred only 1.5% of the time. While the attending physicians erred 14.7%, the resident physicians 13.4%, the nurses 15.9% and the paramedics 17.4%.	Margin of error for dosage not obtained. Repeat weight estimations may have increased accuracy. Patients studied were medically stable.
			Within 5% of weight	Patient 77.5%, Attending physician 33.6&, Resident Physician 34.1% , paramedics 21.5% and nurses 32.9%
			Gender	Female patients underestimated more than men.
			Weight	When patients weight greater than 90kg more likely to under-estimate weight.
Greene S et al 2004 UK	Medical staff were asked to estimate the weight of six patients on a poisons ward. Estimated and actual patient weights were statistically compared.	Cohort observational study	Estimates within 10% Of actual weight	Patient weight was incorrectly estimated by greater than 10% in 61% of individual estimations.	Small number of patients, Only abstract available
Hall WL et al Oct 2004 USA	Providers (n = 33), blinded to study hypothesis and patient data, estimated their own weight as well as the weight of 11-20 patients each. (therefore 494 estimations in total) An independent sample of patients (n = 95) was used to assess biases in patients\' estimation of their own weight.	Cohort study	Weight Estimation	28.1% of the provider estimates were within 5kg. 77.9% of patients were within 5 kg.	Small sample size of providers. Narrow limits of only 5kg. No control for patient age or ethnicity.
			Provider	EM faculty and paramedics more likely to underestimate and the intern and medical student were more likely to overestimate.
			Gender	Female patient more likely to underestimate and male patient more likely to overestimate. Providers more likely to underestimate women than men.
Leary et al Nov 2000 UK	Four experienced intensive care staff (three doctors and one nurse) estimated the weight and measured the height of 30 volunteers and the estimates were compared with accurate reference measurements.	Observational study	No. of observers	3 of the 4 observers were found to be significantly inaccurate using students t test.	Volunteers not patients used. Clinically significant weight difference not defined.
			Pooled estimates	Pooled estimates more accurate then individual estimates.
Fernades et al 1999 Oct Canada	Eleven attending physicians, 26 nurses, and a convenience sample of 117 patients. Patients themselves, attending physicians, and nurses independently estimated the weight of 117 patients. An investigator weighed each patient.	Prospective observational study	Mean Error	Mean error in estimates was 3.1% for patients, 8.4% for nurses, and 8.1% for physicians.	Excluded patients unable to stand. Convenience sample not consecutive.
			< 5% of body weight	32% of nurses, 39% of physicians, and 82% of patients.
			< 10% of body weight	66% of nurses, 66% of physicians, and 97% of patients.
			<15% of body weight	89% of nurses, 84% of physicians, and 99% of patients.
Husin M et al June 2012 Ireland	100 patients attending outpatient endocrinology and cardiology clinics. In the waiting room, patients were asked to self report their weight, height and waist circumference. Weight was measured to the nearest 0.1kg. The doctors were asked to visually estimate the patient’s weight, height, waist circumference and BMI.	Case control study	Patients Weight estimation	49% of patients underestimated their weight by up to 1.5kg, 35% reported accurately t0 1.5 kg and 16% over-reported their weight.	Selection and measurement bias given nature of clinics
			Physician estimates	There were no differences in patient’s self reported weight and doctor’s weight estimation.
Buckley RG et al 2011 USA	330 patients of which 23 were excluded with incomplete data sets were enrolled in the study. Each patient estimated their own weight, doctors and nurses were then asked to makes estimates followed by an investigator who carried out 5 anthropometric measurements and performed a calculation. The patient was then weighed.	Prospective observational study	Estimates of male patients >10 kg inaccurate	Patient were greater than 10 kg inaccurate of own weight 6% of time, while the physicians were inaccurate 38% of patients, the nurses were inaccurate in 31% of patients and the anthropometric model was accurate in 12% of patients.	Relatively small sample size. Unable to stratify got age or ethinicity.
			Estimates of female patients >10 kg inaccurate	Patient were greater than 10 kg inaccurate of own weight 1% of time, while the physicians were inaccurate 32% of patients, the nurses were inaccurate in 27% of patients and the anthropometric model was accurate in 9% of patients
Metcalf AK et al 2009 Nov UK	Doctors, nurses and medical students (n=95) visually estimated the weight of 30 patients. A total of 899 estimated weights were analysed. Patients were also asked to estimate their own weight.	Prospective observational study	Weight Estimation	Only half of all estimations were within 10% of the actual weight.	Abstract only – poster at conference
			Health care professional	No statistically significant difference between health care professionals estimations
			Weight of patient	Linear regression showed that healthcare professionals overestimated thinner patients and underestimated larger patients to a level of statistical significance (p<0.0001).
			Patients vs health care professional	No statistically significant difference between patient and healthcare professional discrepancies (p=0.67).
Goutelle et al Aug 2009	3 observers performed a visual estimation of weight in 71 geriatric patients. Estimated body weights were compared to measured body weights.	Observational Study	Weight estimation	Overall results showed that a three observer panel gave better weight estimates than one or two individuals.	Limited clinical experience, small number of observers.

Comment(s)

Using these articles above if we use estimate of patient weight to within 10% (most commonly used outcome measure) as an outcome measure we can see that doctors are accurate in 57.5% (Range 50 – 66%) of patients, nurses are accurate in 60.9% of patients (Range 50% - 73) and patients were accurate in 90% of patients (79.2 – 97%). This inaccuracy has been shown to have an effect on dosing and outcome in 2 articles (dos reis et al and bruer l et al). A combined estimate of multiple health care professional appear to improve accuracy. The weight of the patient appears to have an effect of the estimate with overweight patients being underestimated and underweight patients being overestimated.

Clinical Bottom Line

Visual estimates by health care providers are very inaccurate. Where possible we should be weighting patients. If not possible to weight patient and estimate needed the patient's own estimate is most accurate at estimating their own weight. Further work is needed to examine which tools are most accurate in estimating weight.

References

1. dos Reis Macedo LG. de Oliveira L. Pintao MC et al Error in body weight estimation leads to inadequate parenteral anticoagulation American Journal of Emergency Medicine 2011 July; 613-7
2. Breuer L. Nowe T. Huttner HB et al Weight approximation in stroke before thrombolysis: the WAIST-Study: a prospective observational \" dose finding\" study Stroke 2010 Dec; 2867-71
3. Maskin LP. Attie S. Setten M. et al Accuracy of weight and height estimation in an intensive care unit Anaesthesia & Intensive Care 2010 Sep; 930-4
4. Lin BW. Yoshida D. Quinn J et al A better way to estimate adult patients\'s weights American Journal of Emergency Medicine 2009 Nov; 1060-4
5. Spicer K. Gibson P. Bloe C et al Weight assessment in cardiac patients: implications for prescription of low molecular weight heparin Postgraduate Medical Journal 2009 Mar; 124-7
6. Lorenz MW. Graf M. Henke C. et al Anthropometric approximation of body weight in unresponsive stroke patients Journal of Neurology, Neurosurgery & Psychiatry 2007 Dec; 1331-6
7. Kahn CA. Oman JA. Rudkin SE. et al Can ED staff accurately estimate the weight of adult patients? American Journal of Emergency Medicine 2007 March; 307-12
8. Bloomfield R. Steel E. MacLennan G et al Accuracy of weight and height estimation in an intensive care unit: Implications for clinical practice and research Critical Care Medicine 2006 Aug; 2153-7
9. Menon S. Kelly AM How accurate is weight estimation in the emergency department? Emergency Medicine Australasia 2005 April; 113-6
10. Corbo J. Canter M. Grinberg D. et al Who should be estimating a patient\'s weight in the emergency department? Academic Emergency Medicine 2005 March ;262-6
11. Anglemyer, BL, Hernandez, C, Brice, JH, et al The Accuracy of Visual Estimation of Body Weight in the ED American Journal of Emergency Medicine Nov 2004; 526-529
12. Greene S. Dargan P. Shin GY et al Doctors and nurses estimation of the weight of patients: A preventable source of systematic error Journal of Toxicology - Clinical Toxicology 2004; 611-5
13. Hall WL 2nd. Larkin GL. Trujillo MJ et al Errors in weight estimation in the emergency department: comparing performance by providers and patients Journal of Emergency Medicine 2004 Oct; 219-24
14. Leary TS. Milner QJ. Niblett DJ The accuracy of the estimation of body weight and height in the intensive care unit European Journal of Anaesthesiology 2000 Nov; 698-703
15. Fernandes CM, Clark C, Price A et al How Accurately do we estimate patients’ weight in emergency department? Canadian Family Physician 1999 Oct; 2373-2376
16. Husin M., Kasim S., Tuthill A Accuracy of visual estimation in diagnosing obese individuals- a blinded study. British Journal of Medical Practitioners June 2012
17. Buckley R.G., Stehman C.R., Dos Santos F.L et al Bedside method to estimate actual body weight in the emergency department. Journal of Emergency Medicine 2012;100-104
18. Metcalf A.K., Barger T., Bennett H et al Incorrect visual estimation of body weight in stroke patients by health care professionals and potential clinical impact on calculating intravenous thrombolysis treatment dose. International Journal of Stroke. December 2009
19. Goutelle S. Bourguignon L. Bertrand-Passeron N et al Visual estimation of patients\' body weight in hospital: the more observers, the better? Pharmacy World & Science. 2009 Aug; 422-5,