An alternative method for measuring workload.

An alternative method for measuring workload The predetermined motion time system we use does a better job of measuring laboratory work than time-study methods with stopwatches. It is faster to apply, more reliable, and more accurate.

So we contended in last month's article. In support of these claims, we will now 1) describe how we arrive at a time standard for a specific laboratory procedure, 2) compare this effort with the CAP workload recording method, and 3) examine the differences between the MTM-UAS system and time study approaches in general.

As we noted previously, developers of the MTM family of systems timed the work activities of many hundreds of individuals, and independent experts then validated the measurements. All we have to do is outline a procedure in terms of the motions it requires and add up the predetermined times. Twenty-nine sets of code letters (AA, AB, AC, PA, PB, etc.) summarize the basic motions, like reaching for something or turning, and degrees of difficulty in performing each of them.

In addition, we have devised our own codes for combinations of motions that add up to a commonly performed task. For example, once we have studied the motions that go into an activity like tube pouring and calculated the time they take, we can assign the entire activity a single code. Thereafter, whenever an analyst develops a time standard for a procedure that includes tube pouring, he or she can simply enter the code for that step into the computer at the appropriate place, and the computer supplies the time value.

A few keystrokes thus take the place of describing all the motions anew and calculating their times again. Such recurring tasks or steps are called standard data. They may consist of just two elements of information or whole pages.

There are two types of standard data. General standard data, which can be used by departments throughout the hospital, have codes that are prefixed with a G. For example, the code for uncapping a pen is G1. Departmental standard data are code-prefixed with a letter signifying the hospital area they are peculiar to: L in the case of the laboratory. If the standard data can be used in only one section of the lab, we add a letter signifying the section: LU for urinalysis, LH for hematology, LS for serology, etc.

Let's see how the UAS system measures a lab procedure. Stat urinalysis including bile is divided into three segments for work measurement purposes--LU1 (pouring urine), LU8 (urine chemistry), and LU9 (urine prep for microscopic analysis). We will explain a few of the steps and codes in each segment.

As Figure I shows, the pouring urine segment actually starts off with accessing the computer for subsequent entry of test results. The notation here is an example of laboratory standard data. A single code, L4, summarizes the 12 steps required to get into the computer system, provide user identification, and identify the procedure. These steps, most of them keystrokes, take 520 TMUs or time measurement units. That translates into 18.7 seconds, since 1 TMU equals 0.036 seconds.

The remaining steps in the pouring urine segment are coded and timed according to the weight of such objects as a drawer and a tube, the distance covered by different motions, the degree of difficulty in getting objects, and how accurately they have to be placed. Figure II reproduces part of the UAS data card from which we draw our codes and times.

Getting a tube is coded UAD2 in Figure I--the U is a computer code denoting the UAS system, and the AD2 is from the data card. As you can see on the card, AD2 means the tube weighs less than 2 lbs.; lies 8 to 20 inches away (as signified by the 2 in the code); is difficult to get, in this case because it is in a bag with other tubes; and once gotten, is placed in an "approximate" area, which means no pressure or precision is involved in putting it there.

One step, pouring, is coded UPT. PT stands for process time, a stopwatch measurement. Timing rather than motion study is better, for some activities requiring little or no movement.

The second segment of Stat urinalysis is LU8, urine chemistry. It begins with the opening of a urine test strip container and removal of a strip to dip in the urine. While waiting for the strip to react with the urine, a technologist can turn to a computer terminal, record the collection method along with the color and appearance of the urine, and use a refractometer for specific gravity. No time units are assigned to these steps. They take place in internal time of 1.4 minutes--the waiting period for a test strip color reaction.

The third segment of Stat urinalysis is LU9, the microscopic examination. In our laboratory, a microscopic is required if the specimen is hazy or cloudy, or positive for protein, white blood cells, nitrite, or red blood cells. Records show that microscopics are done on 50 per cent of our urinalyses, so the TMU value of this segment is halved.

Cultures, required for positive WBCs and nitrite, are performed on 5 per cent of all urinalyses. Within the microscopic procedure are steps connected with earmarking a specimen for culture. Their time value is reduced to the 5 per cent level.

The listing of steps for the microscopic segment runs 35 lines, six more than the pouring urine segment shown in Figure I. But several codes in the microscopic cover standard data, including L29 for refrigerating urine for cultures and L36 for dispensing one drop of urine into a screwtop vial. Without use of standard data codes, the description of the microscopic would be five times as long.

A definitive comparison of UAS-developed standards with those generated by the CAP workload recording method is beyond the scope of this article. Some observations can be made, however.

* Our urinalysis example was based on a Stat procedure. The CAP notes that "by its nature, the CAP workload recording method averages out the effects of batch size and assumes an average Stat load." For that reason, the CAP cannot quantify the true workload impact of a Stat procedure versus a routine one.

* Times for our Stat urinalysis add up to 6.11 minutes (1.11 for pouring urine, 2.44 for urine chemistry, and 2.56 for urine prep for micro). Batch sizes for urine chemistry and urine prep for micro are 5 and 3, respectively. This yields a batch time standard of 3.37 minutes.

To derive a combined Stat/batch standard, we'll multiply our Stat time by 31 per cent, which is the frequency of Stat urinalyses in our hospital, and the tach time by 69 per cent. The results--1.89 and 2.33 minutes, respectively--add up to 4.22 minutes.

We can develop a comparable CAP standard for our hospital. The CAP urinalysis standard without a microscopic is 4 minutes. The microscopic adds 2 minutes to that, but we only perform a micro 50 per cent of the time, so we'll make it 4 + 1. We also perform three tests--the reducing sugar tablet test, urine protein, and urine bilirubin--that the CAP counts separately under code 81004. At the 13 per cent frequency with which we perform these tests, that adds another 0.39 minutes, making a total of 5.39 minutes for the CAP standard--approximately 30 per cent higher than the 4.22 minutes found under our work measurement system.

* The CAP urinalysis standard has the following components that we account for elsewhere, outside the test procedure: initial handling--we have not included timing-in of the specimen or providing the container to outpatients; daily or routine preparation--we have not included quality control or instrument preparation and cleaning; maintenance and repair; solution preparation--we have not included sulfasalicylic acid preparation; glassware washup; and direct technical supervision.

We feel that inclusion of these components in the standard itself would make it less meaningful. Moreover, it is easier to adjust a standard for changes in test procedure when work other than actual testing is dealt with separately. A significant portion of the analysis we have performed on tests would fit the needs of other laboratories with adjustments in time values for distances within the lab and procedural differences. UAS allows labs to develop their own time measurements and not base them on an average hospital lab.

One of the most important characteristics of the MTM systems is that they are method-specific. They force the work measurement analyst's attention toward the method used rather than the time involved. Once the method is established and the motions are assigned to the method, the codes of the motions are entered into the computer, and the time value is automatically determined.

Since the time standard is based only on the motions used, it is hard to question the validity of the study. Either the motions exist, or they do not. This factual basis makes an analyst's results easy to verify.

Precise method descriptions are required. These are extremely helpful in the training of new technologists and in development of job descriptions. In addition, everyone becomes more method-conscious, and method improvements are likelier. It has been found that they majority of cost reductions achieved in most companies are due to method evaluation and improvements.

Another advantage of the MTM systems is that time standards can be estimated before a new procedure is ever adopted or a new instrument is installed. Planning and costing become more accurate.

Finally, the MTM systems, representing the average operator in any work situation anywhere, lead to consistent time standards throughout the organization. This permits comparison between laboratory sections, between the lab and other hospital departments, and between different labs.

On the other hand, standards developed from time study are only approximations at best. They call for considerable judgment by the time study analyst at every step. The actual recording of the time involves the least amount of judgment, but even there a 10 per cent error is recognized by time study experts.

Some of the variable factors that will significantly affect a time study result are: selection of the worker to be studied; conditions under which the work is performed during the study; the manner in which the operation is broken down into parts or elements; the method of reading the stopwatch; the duration of the time study (the worker may get fatigued in an extended study); rating of the worker's performance, which is a common adjustment in time studies, from observed time to "normal" time; the amount of allowance made for personal needs, fatigue, and delay; how the allowance is applied; and the method of computing the time standard for the job from the time data.

As we explained last month, when the cycle time or length of work is at least 5.27 minutes, we are 95 per cent confident that the deviation of our time standard is within plus-or-minus 5 per cent of the true value. As for how long time analyses take, we estimate that the manual UAS system has an application speed of about five to eight times the cycle time, while computerized applications are four to six times the cycle time. Thus it took us 30 minutes to analyze the Stat urinalysis procedure, which has a cycle time of 6 minutes.

To attain accuracy of plus-or-minus 5 per cent with 95 per cent confidence in time studies, it is usually necessary to make a large number of observations because of the variation from one observation to the next. Figure III shows a sample of 20 readings that might be found for the same time study and gives the standard deviation. Figure IV uses a standard statistical formula to determine the number of observations required for plus-or-minus 5 per cent accuracy with 95 per cent confidence, based on the first 20 observations. The answer is 52.

Fifty-two observations of a six-minute procedure like Stat urinalysis total 312 minutes or 5.2 hours. And that doesn't include the time it takes an analyst to evaluate the study, complete the mathematical calculations and statistical reliability checks, and rate the performance of the operator.

Remember, the UAS system did the job in a half-hour. Furthermore, a computer-based work measurement system like ours greatly reduces analyst error. The computer makes many of the decisions, so the analyst avoids mistakes like selecting the wrong time value and adding incorrectly.

Our laboratory used to employ the CAP workload recording method. We switched to our current system because we believed there had to be a better way. Now we are convinced we have found it.