Clinical evaluation of a simple uroflowmeter for categorization of maximum urinary flow rate

Objective: To evaluate the accuracy and diagnostic usefulness of a disposable flowmeter consisting of a plastic funnel with a spout divided into three chambers. Materials and Methods: Men with lower urinary tract symptoms (LUTS) voided sequentially into a standard flowmeter and the funnel device recording maximum flow rate (Q ). The device was precalibrated such that filling of the bottom, middle and top chambers categorized maximum input void flows as <10, 10-15 and > 15 ml s respectively. Subjects who agreed to use the funnel device at home obtained readings of flow category and V void twice daily for seven days. Results: A single office reading in 46 men using the device showed good agreement with standard measurement of Q max for V void > 150 ml (Kappa = 0.68). All 14 men whose void reached the top chamber had standard Q max > 15 ml s (PPV = 100%, NPV = 72%) whilst eight of 12 men whose void remained in the bottom chamber had standard Q max < 10 ml s (PPV = 70%, NPV = 94%). During multiple home use by 14 men the device showed moderate repeatability (Kappa = 0.58) and correctly categorized Q max in comparison to standard measurement for 12 (87%) men. Conclusions: This study suggests that the device has sufficient accuracy and reliability for initial flow rate assessment in men with LUTS. The device can provide a single measurement or alternatively multiple home measurements to categorize men with Q max < 15 ml s .

Measurement of maximum urinary flow rate (Q max ) is men with LUTS but this approach may not be ideal because widely used in the assessment of men complaining of of physiological variation and nonrepresentative V .lower urinary tract symptoms (LUTS).Although Q max Approaches to address these deficiencies include obtaining varies with age and voided volume (V void ), a reduced multiple office readings which may improve diagnostic flow rate can be used clinically to suggest the presence accuracy but is time-consuming and costly [4] or provision of of bladder outlet obstruction (BOO).For clinical home electronic flowmeters which are also expensive and categorization cut-off values have been identified difficult to maintain. [5]Another possibility is home use of a (V ≤ 15 ml s -1 have an disposable uroflowmetry device which enables multiple ) and voided volume max [3] void whereby men with Q max approximate 70% chance of having BOO whilst men with a value > 15 ml s -1 have a 65% chance of not having BOO. [1]Standard uroflowmeters differentiate urine weight change to give a continuous plot of flow rate against time which is smoothed by internal electronic filtering to allow precise (± 5%) measurement of Q max . [2]Most clinicians use a single office measurement of flow rate as part of their assessment of measurements in line with an individual's day-to-day voiding habits and could potentially be used as part of initial assessment of men with LUTS prior to specialist referral. [6]An ideal device would be accurate, simple to use and inexpensive.Prototype devices have been developed but, to our knowledge, are not being routinely used in practice. [6,7]Recently, a simple inexpensive funnel device has been made available which is potentially suitable for repeated measurement of maximum flow rate in the patient's home (Uflow meter, MDTi Ltd, Wolverhampton, UK).We now describe the results of a clinical study which aimed to determine the accuracy and test-retest reliability of the new device in office and home settings with reference to the current standard of a single Q max office-based reading.

The device
The flow device consists of a plastic funnel formed from a cup and a spout divided into three chambers with a 4.6 mm diameter aperture placed at the apex [Figure 1].Fluid poured into the cup will start to fill the funnel as well as flowing out through the aperture.Once inflow (determined by urine flow rate) and outflow (determined by the fixed-diameter aperture and the pressure-head of fluid above it) are equal, a constant maximal fluid level within the funnel will be maintained.With higher filling rates more fluid will be retained in the device

Data analysis
In order to assess the level of agreement between continuous data obtained by the standard uroflowmeter against categorical data obtained using the funnel device we assigned the Q max value obtained by standard uroflowmetry to the appropriate category defined on the funnel device (≤10, >10 -≤15 or >15 ml s -1 ) and then calculated the weighted Kappa statistic (chance corrected correlation coefficient) whereby Kappa > 0.4, > 0.6 and > 0.8 defines moderate, good and excellent agreement respectively.Test-retest reliability of home use of the device was also assessed using the Kappa statistic by comparing all home readings obtained by each and the fluid level will rise in the column provided that the individual to the most frequent flow category (mode) time taken to reach the maximum level (response time) is documented by that subject.The clinical usefulness of multiple substantially less than voiding time.Aperture diameter and home readings using the funnel device was determined by volume of each chamber are calibrated such that filling of the calculating the sensitivity and specificity of the average home bottom, middle and top (including cup) chambers corresponds flow against the reference of a single standard office to input flows of < 10, 10-15 and > 15 ml s -1 respectively to fit measurement for the categories ≤15 and >15 ml s -1 .Differences with the current clinical decision-making.The highest in voided volume were analyzed by paired Student's 't' test chamber reached by urine during the course of a void is with significance level set at P< 0.05.recorded by the patient and categorizes their Q max as being within the range for that chamber.

Subjects approval and with informed consent, we recruited men with
We recruited 46 men with median age 64 (range, 46-82) LUTS attending for standard office uroflowmetry.Each years, of whom 40 (87%) produced two consecutive flows subject performed two sequential voids in a randomized order with voided volume > 150 ml.Most subjects found the funnel either into a standard rotating disc uroflowmeter (Urodyn device easy to use and read whilst five (11%) had difficulty 1000, Medtronic Ltd., Watford, UK) or into the portable due to obesity (n=3) or inability to observe the device and funnel device.The patient held the funnel vertically with the void simultaneously (n=2).spout above a measuring jug placed on the toilet seat lid to measure flow category and V void .For the portable device flow Accuracy category was derived from the highest chamber reached as Figure 2 compares single measurements of Q observed by the patient and verified by the investigator.The the standard uroflowmeter and the funnel device.Men whose printout from the standard flowmeter was manually read and voids remained within the bottom chamber (<10 ml s -1 ; n=12) taken at the highest point of the flow curve discounting had a mean (SD) Q with standard uroflowmetry of 9 (4.0) max max spike artefact and with internal filtering set at 10Hz.ml s -1 , whilst voids that reached the middle chamber (10 -15 Participants in the office study who used the portable device ml s -1 ; n=20) or top chamber/cup (>15 ml s -1 ; n=14) had mean successfully and consented to home use were given a device, (SD) Q of 14 (4.0) ml s -1 and 24 (8.3) ml s -1 respectively.All  14 men with office funnel device readings in the top chamber or cup had Q max > 15 ml s -1 using standard uroflowmetry (PPV = 100 %) whilst eight of the 10 men with standard Q max < 10 ml s -1 were correctly categorized by the device [Table 1].Overall single office measurement of Q max by the funnel device showed good agreement with standard uroflowmetry (Kappa = 0.61).If data from six men with at least one V void < 150 ml were excluded, the agreement level was improved (Kappa = 0.68).The mean (SD) difference between V void for standard uroflowmeter reading and that for the funnel device reading was -17 (157 ml).

Home observations Subjects
A total of 14 men with median age 64 (range, 50-81) years used the device at home and all completed the protocol of 14 Single office flow was assigned to appropriate funnel category and compared to each of the 14 home voids performed by each subject using the funnel device where 0 = same category as office reading, +1 = home reading 1 category higher than office reading, +2 = home reading 2 categories higher than office reading, -1 = home reading 1 category lower than office reading, -2 = home reading 2 categories lower than office reading.The difference in voided volume was calculated as: home V office V void patients reported any difficulty in obtaining recordings using the device at home.

Accuracy
Averaged home readings using the funnel device correctly categorized seven of the nine men with standard Q max > 15 ml s -1 (sensitivity = 78%, specificity = 71%) and all five men with a standard Q max measurement ≤15 ml s -1 (sensitivity = 100%, specificity = 71%; [Figure 3]).For individual home readings the error in categorization compared to the reference office reading was significantly associated with lower V void [Table 2].Test-retest reliability using the funnel device was moderate (Kappa = 0.58).The mean (SD) difference between V void for standard test and mean V void for home uroflowmetry within each individual was -54 (67) ml.
The ageing population and heightened awareness of prostate cancer have increased the number of men with LUTS requesting specialist assessment.This has led many urology practices to set up 'one stop' clinics facilitated by asking men to complete symptom questionnaires and frequency-volume charts prior to the office appointment.The addition of 'home' uroflow measurement to this preassessment would further streamline the process and help decide management options.The novel device assessed in the present study is potentially suited to this use since it appears to offer acceptable accuracy and reliability with ease of patient use at low cost.
For office use our device showed good agreement with ) .
standard measurement, particularly using established diagnostic cut-off values.Most discrepancies occurred due to underestimation by the device of flows in the 15-20 ml s -1 range.This may have been partly due to the tendency for lower V void using the device despite randomization or may reflect the known test-retest variation in the standard measurement of Q max (SD = 2 ml s -1 ). [8,9]e home part of the study was conceived to assess both testretest reliability of the device and predictive value of multiple recordings compared to a single electronic office measurement.The device showed moderate reliability as flow assessment if required.Home flow measurements would also be useful for those patients who are unable to provide a representative void during an office visit.We next intend to establish the minimum number of 'home' flows required to give the best estimate of standard Q max and then compare patient management with and without clinician use of the data generated.

CONCLUSION
This simple inexpensive uroflowmetry device allows multiple estimates of Q max to be made in the home setting.The indicated by the Kappa statistic which may reflect the known accuracy and reliability of the device appears sufficient to with voided volume and time of void [3] or allow categorization around the standard threshold of 15 ml max readings being 'borderline' between two chambers.In s -1 suggesting its usefulness in the preliminary assessment of common with a previous study using a home-based electronic men with LUTS prior to an office visit.flowmeter our home device tended to underestimate a standard office measurement of Q max . [5]Our data suggest that ACKNOWLEDGEMENTS this may be at least partly related to lower home V void and this again is consistent with previous work. [5,10]Encouragingly, Manufacture of prototype devices used in this study was agreement was best for flows ≤ 15 ml s -1 where diagnostic financed by the Business Development Unit, Newcastle decisions will be crucial.If the new device was used at home University and RTC North, a branch of the UK NHS Research as a 'screen' for onward referral for formal uroflowmetry and Development Directorate.Clive Griffiths and Michael underestimation of flow rates > 15 ml s -1 would be of less Drinnan, Department of Medical Physics, Freeman Hospital, concern since this would not reduce test sensitivity for Newcastle upon Tyne gave helpful advice on data analysis detection of flow ≤ 15 ml s -1 .Taken together these findings and constructive comments on the manuscript.We thank suggest that multiple use of our simple uroflowmetry device Philip Harley of Newcastle University for providing Figure 1.does provide a valid and reasonably reliable measure of Q max We thank the men who took part in the study and the staff of and that multiple readings using the device could be of the urodynamic clinic, Freeman Hospital who helped with potential use in the initial assessment of men with LUTS. recruitment.
In terms of simplicity the device tested in the present study is

REFERENCES
intermediate between the very basic 'Streamtest' Cup proposed by Currie [7] and the more complicated multi-exit port device 1. Poulsen A, Schou J, Puggaard L, Torp-Pedersen S, Nordling J. Prostatic of Pel and van Mastrigt. [6]The device showed similar accuracy enlargement, symptomatology and pressure/flow evaluation: using different flow thresholds to the 'Streamtest' but inferior Interrelations in patients with symptomatic BPH.Scand J Urol reliability to the multi-port device which was, however, tested Nephrol Suppl 1994;157:67-73.
on young noncomplaining volunteers with higher V .We widely used flow rate bands together with its manufacture using The addition of flow data to the widely used frequency volume chart completed prior to the initial office visit would greatly aid diagnosis and treatment planning.In the UK and other countries where all men are initially seen by a family care practitioner it could also be used to help select men for specialist referral.Men with most readings in the upper chamber or cup (notional average flow > 15 ml s -1 ) are unlikely to have BOO and alternative explanations for their symptoms should be sought or else they could be monitored in the community.Those with multiple flows in the middle or lower chambers (≤15 ml s -1 ) could be selected for urology referral and a formal al.: Evaluation of simple home uroflowmeter simple instruction sheet to record flows in a similar manner twice daily for seven days noting Q max category and V void on each occasion.

Figure 1 :
Figure 1: Photographs of the device showing a, the 3-chamber design and b, the 4.6 mm aperture

Figure 2 :
Figure 2: Scatter plot showing the relationship between maximum flow ) readings obtained in the office using the standard uroflowmetry (vertical axis) and those obtained with the funnel device (horizontal axis).The mean of the single standard office measurements for subjects voiding within each funnel category is indicated by the l i c a t i o n s ( w w w .m e d k n o w .c o m 2. Schafer W, Abrams P, Liao L, Mattiasson A, Pesce F, Spangberg A, et al. void Good urodynamic practices: Uroflowmetry, filling cystometry and feel that the ability of the funnel device to categorize men into pressure-flow studies.Neurourol Urodynamics 2002;21:261-74.

3.
Golomb J, Lindner A, Siegel Y, Korczak D. Variability and variation of Q a single plastic molding combines low cost [estimated at $8.60 (≤ 5; £ 7.30) per unit] and acceptable accuracy making it suitable for disposable use prior to an office visit for men complaining of LUTS.