Abstract
Current literature does not support routine testing for hereditary and acquired thrombophilia disorders in the inpatient setting. Testing in the acute setting rarely changes patient management or could lead to patient mismanagement. Despite prior educational interventions, continued overuse of inpatient testing warrants further quality improvement measures. A hard-stop best practice advisory pop-up was implemented in the electronic medical record in a multicenter academic hospital system to provide clinicians guidance on the appropriate use of thrombophilia testing at the point of care. Pre- and postintervention retrospective data were collected to assess clinical features before and after implementation. Before the intervention, 271 patients underwent inpatient hypercoagulability testing; after the intervention, 238 patients underwent inpatient hypercoagulability testing. The total number of labs ordered per patient decreased from 1185 to 910, a 13% reduction (P = 0.003). Overall, there was a savings of $23,597 in total direct cost and $123,153 in total charges when comparing the 6-month timeframes before and after the intervention (P < 0.01). Although this study found only mild reductions in thrombophilia testing, it presents a new means of providing point-of-care intervention and education for hypercoagulability testing in the inpatient setting.
Keywords: Benign hematology, hypercoagulability testing, quality control, thrombophilia testing
Acquired and inherited forms of thrombophilia represent a variety of conditions that confer an increased risk of thrombosis.1 Unprovoked venous thromboembolic events, thrombotic events in young patients, or unusual sites of thrombosis (i.e., splanchnic, mesenteric, portal vein, cerebral venous sinus, etc.) are common scenarios where a hypercoagulable state may be suspected. Antiphospholipid antibody syndrome (APS) is an acquired autoimmune condition that is thought to lead to abnormal activation of the coagulation cascade.2 Although a variety of serum tests are available to evaluate these conditions, studies suggest that test results poorly predict thromboembolism recurrence.3,4 Literature demonstrates that inpatient workup of hereditary thrombophilia rarely changes patient management.5,6 For these reasons, guidelines on thrombophilia testing discourage hereditary thrombophilia workup for provoked thrombotic events in the inpatient setting.7–9 In instances of recurrent or unprovoked thrombosis, however, there may be some utility to inpatient thrombophilia testing. Hypercoagulability testing panels are costly, with prices ranging from $1100 to $2400.10,11
To reduce inpatient thrombophilia testing, several educational interventions were previously implemented.12,13 An electronic medical record (EMR) intervention in the form of a patient advisory pop-up was tested, but did not result in a statistically significant reduction in inpatient testing.14 This may be attributed to the pop-up advisory not being a hard stop, which allowed providers to easily bypass it. The current quality improvement project involved the implementation of a hard stop pop-up advisory offering education and requiring providers to disclose their reasoning for ordering the test before signing the order.
METHODS
The incidence of inpatient hereditary and acquired thrombophilia testing in patients at two academic hospitals (>1600 hospital beds combined) from the same multicenter institution was assessed via retrospective EMR (Epic) chart review for a 6-month period (December 1, 2019, to May 31, 2020). Any adult patient who underwent thrombophilia testing was evaluated after a query was made of specific laboratory tests ordered. The search included only hospitalized patients; patients in the emergency department or urgent care setting were excluded. The tests queried were those associated with the most well-described hereditary hypercoagulable disorders, including factor V Leiden, prothrombin gene testing, protein C, protein S, and antithrombin deficiencies. Additionally, antiphospholipid antibodies (such as anticardiolipin and anti-beta 2 glycoprotein I antibodies) and the lupus anticoagulant panel (which includes prothrombin time, the LA-sensitive prothrombin time, and dilute Russell viper venom test) were evaluated. The indications of testing were assessed via retrospective chart review and categorized as unprovoked, having suggestive clinical characteristics of a hypercoagulable state (age <50 years, unusual site of venous thrombosis, family history of thrombosis, or recurrent thrombosis), or having a weak provocation factor (the use of oral contraceptives, decreased mobility, or minor procedure/surgery). Other features assessed included test result status (marked positive or negative). A positive result did not mean that a true disorder was diagnosed but merely that a laboratory abnormality was found that may suggest the presence of a hypercoagulability disorder. The departments associated with the ordering provider and whether a hematology consult was placed were noted. The anticoagulation status of each patient was also recorded. Another variable recorded was whether the patient had an acute thrombotic event (arterial or venous) within the last 30 days at the time of testing.
After 6 months of baseline data acquisition and review, an intervention was created and disseminated at the same academic hospitals on December 1, 2020. An educational advisory pop-up was created to appear when the individual thrombophilia tests were requested in the EMR (Epic). This electronic order intervention was a hard stop pop-up best practice advisory that was triggered when any of the above-mentioned tests were ordered in the hospital setting (Figure 1). The pop-up provided brief education and could only be bypassed if the ordering provider selected an option indicating that the test would impact inpatient patient management or was requested by the hematology consult service. Additionally, the intervention involved the removal of an inpatient hypercoagulability order set that contained the above-mentioned tests.
Figure 1.
Best practice advisory pop-up generated when a hypercoagulable workup was ordered on the inpatient electronic orders (demonstration with a false patient name). The hard stop could be bypassed if the ordering provider indicated that the order would either impact inpatient management or was requested by the hematology consult service.
Six months of postintervention data were collected (December 1, 2020, to May 31, 2021) to assess for statistically significant changes in testing frequency (primary outcome), characteristics, and associated healthcare costs. Patient consent was not required in this study, as patient-identifying information was completely hidden from all reporting. An institutional review board waiver was obtained prior to initiating this project.
Continuous variables were summarized by mean and standard deviation or median and interquartile range whenever appropriate. Categorical variables were summarized by frequency and percentages. Comparison of patient characteristics and outcomes between indication groups were assessed by chi-square test for categorical variables and independent sample t test/Kruskall-Wallis test for continuous variables. Association between indication and positive test result for different laboratory tests was assessed using the chi-square model. Number of lab tests per patient was modeled with the Poisson model. Cost and charges were modeled with the Gaussian model, and indicators of active thrombosis or active coagulation were modeled with logistic regression models. Independent risk factors that were adjusted for included age, sex, ordering department, indicator hematology consultation, and positive test status. Statistical analysis was conducted using R Statistical Software. A P value < 0.05 was considered statistically significant.
RESULTS
On retrospective chart review, 271 patients underwent a hypercoagulability workup in the inpatient setting during the 6-month preintervention period. There was approximately a 6-month delay in implementing the intervention after the acquisition of baseline data due to hospital administrative and information technology processing and guidance prior to approval. After implementing the electronic orders intervention in early December 2020, 238 patients were identified who underwent hypercoagulability workup during the 6-month postintervention period. This represented a 12% reduction in the number of patients tested, but the study was not powered to evaluate the statistical significance of this difference. There was a statistically significant reduction in the total number of laboratory tests ordered, from 1185 in the preintervention group to 910 in the postintervention group (Table 1), which amounts to a 13% decrease (P = 0.003). The individual tests were evaluated as well, demonstrating a statistically significant reduction in some of the individual tests ordered but also an increased incidence of orders placed for APS workup. It is difficult to assess how the COVID-19 pandemic and resulting hospital surges may have impacted the number of tests ordered and number of patients tested.
Table 1.
Number of laboratory tests at baseline and after the intervention
| Baseline (N = 271) | Postintervention (N = 238) | Change from baseline | P valuea | |
|---|---|---|---|---|
| Total number of labs (rate per patient) | 1,185 (4.37) | 910 (3.82) | −13% | 0.003 |
| Antiphospholipid antibodiesb | 523 (44.1%) | 497 (54.6%) | +24% | 0.22 |
| Protein S activity | 96 (8.1%) | 77 (8.5%) | +5% | 0.61 |
| Antithrombin III activity | 104 (8.8%) | 50 (5.5%) | −38% | <0.001 |
| Protein C activity | 54 (4.6%) | 22 (2.4%) | −48% | 0.002 |
| Prothrombin gene mutation | 94 (7.9%) | 46 (5.1%) | −35% | 0.001 |
| Factor V Leiden | 101 (8.5%) | 53 (5.8%) | −32% | 0.002 |
| Lupus anticoagulant panelc | 213 (18.0%) | 165 (18.1%) | +0.1% | 0.24 |
P value based on Poisson test.
Antiphospholipid antibodies comprised anticardiolipin and anti-beta 2 glycoprotein I antibodies.
The lupus anticoagulant panel included a prothrombin time test, the LA-sensitive prothrombin time test, and dilute Russell viper venom test.
Table 2 describes and compares patients’ demographic and clinical characteristics. There was no statistically significant difference in terms of demographics, ordering departments, or percentage of positive tests in the pre- or postintervention populations. More women than men were tested, with 154 women (56.8%) in the preintervention group and 143 women (60.1%) in the postintervention group. The average patient age at testing was 49.2 years in the preintervention data and 48.6 years in the postintervention data (P = 0.69). The ordering provider departments included hematology, hospitalist medicine, surgery, and other subspecialists. Hematology was consulted 101 times (37.3%) in the preintervention group and 99 times (42.7%) in the postintervention group (P = 0.22). Only 58 tests (22.3%) in the preintervention group and 65 tests (28.3%) in the postintervention group yielded a positive result (P = 0.13).
Table 2.
Demographic and clinical features of the pre- and postintervention patient populations
| Baseline/preintervention (N = 271) | Postintervention (N = 238) | P value | |
|---|---|---|---|
| Age (years): mean (SD) | 49.2 (16.5) | 48.6 (18.5) | 0.69a |
| Sex | 0.46b | ||
| Female | 154 (56.8%) | 143 (60.1%) | |
| Male | 117 (43.2%) | 95 (39.9%) | |
| Ordering department/provider | 0.44b | ||
| Hematology | 69 (25.5%) | 66 (28.4%) | |
| Hospitalist | 88 (32.5%) | 60 (25.9%) | |
| Other subspecialties | 82 (30.3%) | 75 (32.3%) | |
| Surgery | 32 (11.8%) | 31 (13.4%) | |
| Hematology consulted | 101 (37.3%) | 99 (42.7%) | 0.22b |
| Positive result | 58 (22.3%) | 65 (28.3%) | 0.13b |
P value based on Student’s t test.
P value based on Pearson’s chi-squared test.
Indications for testing were assessed via retrospective chart review. Around 6% of cases of thrombophilia workup were performed for unprovoked thrombosis (Table 3). Hypercoagulability testing for suggestive clinical characteristics of a hypercoagulable state, including age under 50, unusual site of venous thrombosis, family history of thrombosis, or recurrent thrombosis, composed 23.2% of testing indications in the preintervention population vs 22.8% in the postintervention population. Testing for thrombosis which occurred under weak provocation (due to the use of oral contraceptives, decreased mobility, or minor procedures/surgery) occurred for 40.6% in the preintervention population vs 35.3% in the postintervention population. Other indications for testing composed 31% of the preintervention tests and 36.2% of the postintervention testing. Within the “other” category, some of the most common indications included testing in the setting of unexpected fetal demise, an organ transplant rejection, or COVID-19 infections with the purpose of assessing risk factors for a hypercoagulable state. There was no statistically significant change in the indication for testing before and after the intervention (P = 0.58).
Table 3.
Indications for the hypercoagulable workup and outcomes of testing
| Variable | Baseline (N = 271) | Postintervention (N = 238) | P value |
|---|---|---|---|
| Indication | 0.58a | ||
| 1. Unprovoked | 14 (5.2%) | 13 (5.6%) | |
| 2. Suggestive clinical characteristicsd | 63 (23.2%) | 53 (22.8%) | |
| 3. Weak provocatione | 110 (40.6%) | 82 (35.3%) | |
| 4. Other indicationsf | 84 (31.0%) | 84 (36.2%) | |
| On anticoagulation during testing | 188 (69.4%) | 157 (67.7%) | 0.68a |
| Active thrombosis when tested (first 30 days) | 167 (61.6%) | 127 (54.7%) | 0.12a |
| Number of labs ordered per patient | 0.16b | ||
| Median (Q1, Q3) | 4.0 (2.0, 6.0) | 3.0 (3.0, 4.8) | |
| Total direct cost per patient | <0.01c | ||
| Mean (SD) | $191.6 ($133.90) | $119.00 ($80.70) | |
| Sum | $51,916.90 | $28,320.00 | |
| Total charges per patient | <0.01c | ||
| Mean (SD) | $1726.10 ($1128.80) | $1448.00 ($926.70) | |
| Sum | $467,779.60 | $344,626.20 |
P value based on Student’s t test.
P value based on Pearson’s chi-squared test.
P value based on Kruskall-Wallis test.
Age under 50, unusual site of venous thrombosis, family history of thrombosis, or recurrent thrombosis.
Resulting from use of oral contraceptives, decreased mobility, or minor procedures/surgery.
Examples include testing in the setting of unexpected fetal demise, organ transplant rejection, or COVID-19 infections with the purpose of assessing risk factors for a hypercoagulable state.
Of the patients evaluated, 69.4% of the preintervention population and 67.7% of the postintervention population were on anticoagulation during testing (P = 0.68). Of the preintervention patients tested, 61.6% had an active thrombosis (either venous or arterial) and 54.7% had an active thrombosis among the postintervention population (P = 0.12). There was a significant decrease in average direct cost per patient by $26 (P < 0.0001) and charges by $167 (P < 0.0001) (Figure 2). Overall, there was a savings of $23,597 in total direct costs and a savings of $123,153 in total charges when comparing the 6-month time periods (P < 0.01).
Figure 2.
Changes in (a) direct cost and (b) charges per patient (P < 0.0001).
DISCUSSION
Hypercoagulability testing in the acute setting produces results that are difficult to interpret. Due to inflammation from acute thrombosis or other conditions such as pregnancy, there can be acquired or transiently low levels of clotting factors detected including protein C, protein S, and antithrombin III. This would not indicate an inherited disorder, even though the test results would be abnormally low in both settings. Therefore, utilization of the above-mentioned tests in the acute setting can lead to the inaccurate diagnosis of an inherited condition. However, this issue does not apply to tests with more specificity for inherited thrombophilia, such as Factor V Leiden or prothrombin gene mutation. Additionally, active treatment of a patient on anticoagulation can affect the accuracy of concurrent APS testing.15 Inappropriate testing can increase the risk of patient harm when a test is falsely positive, leading to long-term anticoagulation and unnecessarily increased risk of bleeding. Falsely negative test results can lead to a false reassurance to a patient who may be at risk for recurrent thrombosis despite undergoing thrombophilia workup.
This quality improvement initiative set out to reduce the number of inpatient thrombophilia tests ordered. There was a reduction in the frequency of thrombophilia testing ordered and in direct patient cost and charges during a 6-month period. This does not include the presumable improvement in cost to the patient and the health care system by avoiding repeat testing in the outpatient setting and avoiding long-term anticoagulation use in a patient who received inaccurately interpreted lab tests.
As described in the results, 69% of patients were on anticoagulation during testing among the baseline population and 67% of patients were on anticoagulation among the postintervention population (P = 0.68). On further data analysis, the tests ordered in these patients included antithrombin, protein C level, and protein S level, which should not be checked during an active thrombosis event, indicating inappropriate testing. Around 30% to 40% of patients tested during the pre- and postintervention period were not on anticoagulation and did not have an active thrombus (arterial or venous). On further chart review of these patients, many were patients infected with COVID-19 and undergoing workup to determine thrombosis risk. It is unclear how results of this testing would impact patient management in the hospital, as this is a context with limited supportive literature.
Approximately 35% of patients studied in this intervention had a workup for acute cerebrovascular events. One of the questions raised is whether unexpected and unusual arterial thrombotic events such as myocardial infarction or cerebrovascular insults in a young patient warrant inpatient thrombophilia workup, since much of the literature reviewed so far is in reference to venous thromboembolism. It has been demonstrated that hereditary conditions still present minimal to modest risk for arterial thrombosis.12,14 Acquired thrombophilia such as APS has been described in correlation with arterial thrombosis risk, and workup may be desired to assess the need for warfarin over other anticoagulants in select situations. Nonetheless, testing will need to be repeated in the outpatient setting before any diagnosis can be confirmed due to higher rates of false test results during an acute presentation.16,17
Although there was a statistically significant decrease in many of the individual tests ordered, there was an increased incidence of orders placed for APS workup. It is suspected that this may be related to testing being performed in the setting of COVID-19 infection. Likewise, there are reports of increased thrombotic risk in patients with APS on direct oral anticoagulants, and this may further explain the higher utilization of these tests in the inpatient setting.18–21 Of note, the results of this study demonstrated that around 30% of hypercoagulability workup performed pre- and postintervention was under the direction of the hematology department. After further chart review, this was found to be related to one hematologist’s practice preferences. This highlights the need for continued education among providers, including hematologists and medical oncologists.
Admittedly, the COVID-19 pandemic presents a confounding factor to this study. Further limitations of this study include its retrospective nature. Without randomization of patients and direct observation of patient care, our conclusions are less substantiated. Further exploratory analysis would be helpful to better understand the results of this study.
This study presents a method aimed at reducing unnecessary and costly inpatient thrombophilia testing working under the assumption that most inpatient thrombophilia testing is unnecessary or more appropriate in the outpatient setting, except for cases involving recurrent or unprovoked thrombosis. A goal of reducing the incidence of unwarranted thrombophilia testing in the inpatient setting overall is a presumed surrogate for an improvement in patient care and better resource utilization. Analysis of the postintervention data demonstrated a reduction in the number of tests ordered and reduction in healthcare expenditure. There was also a reduction in the number of patients tested overall, but the statistical significance of this could not be assessed with the data obtained. It is possible that the number of tests ordered would have been more significantly reduced had the COVID-19 pandemic not been a confounding variable. Given the retrospective nature of this analysis, further study is needed to prove this trend has accurately resulted from the EMR hard stop intervention.
Educational interventions targeting excessive thrombophilia workup have been utilized but follow-up was brief, and interventions often occurred only for a limited duration. This study sought to address this issue by proposing a method of ongoing education at the point of care through an electronic intervention for every provider placing an order for inpatient thrombophilia workup. Appropriate utilization of educational hard stops in the EMR presents a strategy to improve testing overuse or inappropriate testing patterns in the hospital setting.
References
- 1.Keeling D, Mackie I, Moore GW, Greer IA, Greaves M, British Committee for Standards in Haematology. Guidelines on the investigation and management of antiphospholipid syndrome. Br J Haematol. 2012;157(1):47–58. doi: 10.1111/j.1365-2141.2012.09037. [DOI] [PubMed] [Google Scholar]
- 2.Harper BE, Wills R, Pierangeli SS.. Pathophysiological mechanisms in antiphospholipid syndrome. Int J Clin Rheumtol. 2011;6(2):157–171. doi: 10.2217/ijr.11.9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Middeldorp S. Evidence-based approach to thrombophilia testing. J Thromb Thrombolysis. 2011;31(3):275–281. doi: 10.1007/s11239-011-0572-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Christiansen SC, Cannegieter SC, Koster T, Vandenbroucke JP, Rosendaal FR.. Thrombophilia, clinical factors, and recurrent venous thrombotic events. JAMA. 2005;293(19):2352–2361. doi: 10.1001/jama.293.19.2352. [DOI] [PubMed] [Google Scholar]
- 5.Kearon C. Influence of hereditary or acquired thrombophilias on the treatment of venous thromboembolism. Curr Opin Hematol. 2012;19(5):363–370. doi: 10.1097/MOH.0b013e328356745b. [DOI] [PubMed] [Google Scholar]
- 6.Merriman L, Greaves M.. Testing for thrombophilia: an evidence-based approach. Postgrad Med J. 2006;82(973):699–704. doi: 10.1136/pgmj.2006.048090. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Hicks LK, Bering H, Carson KR, et al. The ASH Choosing Wisely campaign: five hematologic tests and treatments to question. Blood. 2013;122(24):3879–3883. doi: 10.1182/blood-2013-07-518423. [DOI] [PubMed] [Google Scholar]
- 8.Baglin T, Gray E, Greaves M, et al; British Committee for Standards in Haematology . Clinical guidelines for testing for heritable thrombophilia. Br J Haematol. 2010;149(2):209–220. doi: 10.1111/j.1365-2141.2009.08022.x. [DOI] [PubMed] [Google Scholar]
- 9.Stevens SM, Woller SC, Bauer KA, et al. Guidance for the evaluation and treatment of hereditary and acquired thrombophilia. J Thromb Thrombolysis. 2016;41(1):154–164. doi: 10.1007/s11239-015-1316-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Gupta A, Patel P, Anwar R, Villanueva D, Vasudevan V, Guevara E.. Hypercoagulable workup in a community hospital setting: to test or not to test; that is the question. J Community Hosp Intern Med Perspect. 2019;9(5):392–396. doi: 10.1080/20009666.2019.1655627. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Somma J, Sussman II, Rand JH.. An evaluation of thrombophilia screening in an urban tertiary care medical center: A “real world” experience. Am J Clin Pathol. 2006;126(1):120–127. doi: 10.1309/KV06-32LJ-8EDM-EWQT. [DOI] [PubMed] [Google Scholar]
- 12.Morris JG, Singh S, Fisher M.. Testing for inherited thrombophilias in arterial stroke: can it cause more harm than good? Stroke. 2010;41(12):2985–2990. doi: 10.1161/STROKEAHA.110.595199. [DOI] [PubMed] [Google Scholar]
- 13.Bereczky Z, Balogh L, Bagoly Z.. Inherited thrombophilia and the risk of myocardial infarction: current evidence and uncertainties. Kardiol Pol. 2019;77(4):419–429. doi: 10.33963/KP.14804. [DOI] [PubMed] [Google Scholar]
- 14.Boekholdt SM, Kramer MH.. Arterial thrombosis and the role of thrombophilia. Semin Thromb Hemost. 2007;33(6):588–596. doi: 10.1055/s-2007-985755. [DOI] [PubMed] [Google Scholar]
- 15.Connors JM. Thrombophilia testing and venous thrombosis. N Engl J Med. 2017;377(12):1177–1187. doi: 10.1056/NEJMra1700365. [DOI] [PubMed] [Google Scholar]
- 16.Giannakopoulos B, Passam F, Ioannou Y, Krilis SA.. How we diagnose the antiphospholipid syndrome. Blood. 2009;113(5):985–994. doi: 10.1182/blood-2007-12-129627. [DOI] [PubMed] [Google Scholar]
- 17.Shen YM, Tsai J, Taiwo E, et al. Analysis of thrombophilia test ordering practices at an academic center: a proposal for appropriate testing to reduce harm and cost. PLoS One. 2016;11(5):e0155326. doi: 10.1371/journal.pone.0155326. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Ordi-Ros J, Sáez-Comet L, Pérez-Conesa M, et al. Rivaroxaban versus vitamin K antagonist in antiphospholipid syndrome: A randomized noninferiority trial. Ann Intern Med. 2019;171(10):685–694. doi: 10.7326/M19-0291. [DOI] [PubMed] [Google Scholar]
- 19.Woller SC, Stevens SM, Kaplan D, et al. Apixaban compared with warfarin to prevent thrombosis in thrombotic antiphospholipid syndrome: a randomized trial. Blood Adv. 2022;6(6):1661–1670. doi: 10.1182/bloodadvances.2021005808. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Sato T, Nakamura H, Fujieda Y, et al. Factor Xa inhibitors for preventing recurrent thrombosis in patients with antiphospholipid syndrome: a longitudinal cohort study. Lupus. 2019;28(13):1577–1582. doi: 10.1177/0961203319881200. [DOI] [PubMed] [Google Scholar]
- 21.Pengo V, Denas G, Zoppellaro G, et al. Rivaroxaban vs warfarin in high-risk patients with antiphospholipid syndrome. Blood. 2018;132(13):1365–1371. doi: 10.1182/blood-2018-04-848333. [DOI] [PubMed] [Google Scholar]