Phlebotomy Staffing: A College of American Pathologists Q-Probes Study of 42 Clinical Laboratories

For the purposes of ensuring the comparability of data emanating from all 3 studies, we defined key terms, which are described in Table 1.

Subscribers to this College of American Pathologists Q-Probes study submitted data on the numbers of staffing personnel and volumes of phlebotomies performed in their laboratories along with various phlebotomy practice characteristics of their overall laboratory, selected inpatient sites, and selected outpatient sites. Of the 42 study participants, 39 were from the United States, 1 was from Canada, and 2 were from Saudi Arabia. Additional participant demographics are summarized in Table 2.

We instructed participants to tally onto standardized data collection forms that we provided the following data in up to 2 hospital inpatient and up to 3 outpatient phlebotomy sites during a full Monday through Friday work week: numbers of inpatients and outpatients on whom they performed phlebotomy tasks and the number of full-time equivalent (FTE) staff hours their phlebotomists devoted to performing venipunctures and nonvenipuncture phlebotomy tasks (see definitions). We excluded from all phlebotomy counts specimens that were obtained by nonlaboratory personnel. We also asked participants to provide the number of billable tests obtained for all inpatients and outpatients that were obtained by laboratory phlebotomy and the number of phlebotomy FTEs (head count) for the previous 1-year period.

From these data, we calculated the performance indicators that are summarized in Table 3. So that participants might evaluate their phlebotomy staffing ratios within their peer groups for selected inpatient or outpatient sites, we calculated the number of phlebotomies per FTE and the ratios of phlebotomies per FTE for laboratories performing more and fewer than/equal to 1350 inpatient phlebotomies and more and fewer than/equal to 340 outpatient phlebotomies per week (Table 4). We established these peer groups to allow study participants to compare their performance with that of participating institutions with similar workloads. We established these peer groups by determining testing volume breakpoints, which we computed using an average linkage clustering algorithm that defines groups that minimize variances within clusters using the average distance between pairs of observations. We then defined testing volume peer-group breakpoints by applying test volume thresholds that separated the clusters while maintaining adequate peer-group sample sizes.

In order to investigate conditions that might influence productivity, participants completed a general questionnaire surveying their laboratory phlebotomy practices. We tested the associations of performance indicators with several practice variables, including types of laboratory services and staff duties, percentage of phlebotomies performed by laboratory staff, percentage of pediatric patients, presence of roving phlebotomist to assist at site, several barriers to performing phlebotomies, presence of phlebotomy training, and required scheduling of outpatient appointments.

For the overall laboratory performance indicator (performance indicator 1), we tested 1 laboratory practice variable: whether patient satisfaction surveys for phlebotomy were distributed for both inpatients and outpatients (yes—always, yes—sometimes, or no).

For the inpatient performance indicator (performance indicator 2) we tested 4 inpatient site practice variables: categorized number of services in inpatient phlebotomy; categorized percentage of inpatient phlebotomies done by laboratory staff; use of a roving phlebotomist to assist at site (yes always, yes sometimes, or no); and up to 5 barriers encountered by staff when performing phlebotomies, as described in Table 1.

For the 3 outpatient phlebotomies performance indicators (performance indicators 3–5), we tested associations of performance indicators with the following outpatient site practice variables: categorized number of nonphlebotomy staff duties phlebotomists performed (frequently or occasionally); categorized number of staff barriers in outpatient phlebotomy (Table 1) (frequently or occasionally); the percentage of pediatric outpatients of all patients; whether or not phlebotomists were trained on-site (yes or no); and whether or not phlebotomy services were scheduled by appointment (yes or no). Finally, we stratified the number of phlebotomies per phlebotomist FTE (performance indicator 3) by 3 equally chosen patient wait time intervals: 0 to 5 minutes, 6 to 10 minutes, and 11 minutes or more. Median wait time was calculated from the reported average time span (minutes) spent by patients from laboratory registration until their venipuncture samples were obtained.

All statistical analyses were completed using SAS (Cary, North Carolina). For all performance indicators, the distributions were non-Gaussian, so nonparametric statistical tests were used, either Wilcoxon rank sum or Kruskal-Wallis tests. Significant associations were identified for P values < .05.

This study was conducted during the third quarter of 2017. Forty-two institutions participated in this study, providing eligible results for 40 selected inpatient sites and 70 selected outpatient sites. Because not all participants provided both valid numerators and denominators for each performance indicator, the N for each performance indicator varied.

Table 4 summarizes the percentile distributions of performance indicators. Of the participants who provided us valid data, between the 10th and 90th percentiles, billable tests/phlebotomy FTE spanned 23-fold; inpatient phlebotomies per phlebotomy FTE for sites performing more and fewer than 1350 phlebotomies/wk spanned 3.3- to 11.8-fold, respectively; and outpatient phlebotomies per phlebotomy FTE for sites performing more and fewer than 340 phlebotomies/wk spanned 3.9- and 5.0-fold, respectively. The greater the number of weekly phlebotomies, the greater the ratios of phlebotomies to FTEs. The central 80% of average outpatient wait times was 4 to 18 minutes, with a median average wait time of 8 minutes. Table 5 shows the percentile distributions of institutional staffing levels and billable testing volumes reported for the calendar year prior to the study. Between the 10th and 90th percentiles of sites, the number of total inpatient and outpatient billable tests obtained by laboratory phlebotomists spanned greater than 142-fold and 30-fold, respectively, and the number of total employed phlebotomy FTEs spanned 11-fold.

None of the performance indicators were associated with the practice variables that we chose to test. Outpatient productivity measured as the number of phlebotomies per outpatient phlebotomy FTE was not significantly associated with the 3 categorized levels of patient wait times.

Table 6 shows the numbers of laboratories that did and did not distribute satisfaction surveys to patients. Thirteen of 34 participants (38.2%) indicated that they did not survey inpatients regarding satisfaction with phlebotomy services, and 6 of 38 (15.8%) indicated that they did not survey outpatient satisfaction with phlebotomy services. The distribution of satisfaction surveys was not associated with differences in performance indicators.

We attempted to determine national normative ranges of phlebotomy staffing levels and laboratory practices that might influence those ranges. We believed that knowledge of these data might assist laboratory managers in calculating their phlebotomy staffing needs.

The distribution of phlebotomy staffing performance indicators among the laboratories participating in this study varied widely even among those groups performing similar volumes of tests (Table 4). It is tempting to explain these ranges of performance ratios on the basis of variations in economy of scale and greater efficiencies among some laboratories. However, it may not be that simple. Lower performance ratios are not necessarily synonymous with greater productivity. The parameters of local practice environments and not national normative value ranges may be the metrics that inform and determine phlebotomy staffing requirements. As an example, some laboratories more than others may serve clinicians who order a large number of timed tests such as those required for drug and glucose monitoring, blood cultures, and high-priority tests requiring immediate results such as troponin levels. Some clinicians may require expedient reviews of test results, especially when meeting preoperative and postoperative deadlines, or have other unique service demands. Laboratory managers may need to overstaff their phlebotomy departments to meet these requirements.

Some laboratories more than others may require phlebotomists to execute a large number of nonphlebotomy tasks other than performing venipunctures. Phlebotomists may be required to answer telephone calls, register patients, enter data into laboratory computers, perform electrocardiograms, etc. The more time they spend on these nonphlebotomy tasks, the less time they have to perform venipunctures, and in some institutions this may lower performance ratios. Rather than representing consequences of reduced productivity, diverting venipuncture responsibilities to other tasks may be a purposeful effort that frees more-specialized and perhaps higher-paid personnel to perform other tasks that require greater levels of skill. Reducing the number of phlebotomists performing these tasks might improve phlebotomy performance ratios but do so at the expense of increasing laboratory overhead. Laboratory managers must allocate staff in a manner that best serves their entire health care community.

We were unable to demonstrate that longer outpatient phlebotomy waiting times were indicators of lower productivity as measured by the number of outpatient phlebotomies performed per FTE outpatient phlebotomist. The reason for this lack of association is not clear but may be related to our relatively small sample sizes in each waiting time interval.

Only 12 of 34 (35.3%) and 19 of 38 laboratories (50.0%) routinely surveyed inpatient and outpatient satisfaction, respectively (Table 6). Satisfaction with laboratory services may reflect timeliness of laboratory testing,^11–17  which in turn may depend on the numbers of phlebotomy staff required to deliver specimens to laboratories. Some laboratories may need to augment phlebotomy staffing levels in order to meet the demands of customer satisfaction, necessarily lowering performance ratios. This study was not designed to associate performance measures with levels of customer satisfaction with laboratory services.

None of the performance indicators were associated with the practice variables or with the peer groups we chose to test. This did not surprise us. In line with our observations stated above, characteristics such as the diversity of patient needs, technologies, degree of laboratory automation, turnaround time requirements, demands of the medical communities, availability of labor, and testing requirements, about none of which we inquired, may explain our inability to statistically link these associations. The relatively small sample sizes of laboratories sharing similar practices and comprising peer groups may have also compromised our ability to establish statistically significant associations.

We did not investigate whether national normative phlebotomy staffing rates might assist laboratory managers in determining staffing requirements in start-up laboratories or in gauging staffing efficiencies among several laboratories operating under similar conditions in single institutions. In such instances, laboratory managers may find value in using the performance measures we propose as starting points from which to track their success in implementing activities designed to improve productivity. We caution laboratories to avoid implementing improvements in a vacuum. Any efforts to improve productivity must incorporate outcome measures that simultaneously gauge their impact on quality and stakeholder satisfaction.