SEARCH

SEARCH BY CITATION

Keywords:

  • Venous thromboembolism;
  • validity;
  • deep vein thrombosis;
  • pulmonary embolism

ABSTRACT

  1. Top of page
  2. ABSTRACT
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. ACKNOWLEDGEMENTS
  9. REFERENCES

Background

Venous thromboembolism (VTE) is a serious complication. Large claims databases can potentially identify the effects that medications have on VTE. The purpose of this study is to evaluate the evidence supporting the validity of VTE codes.

Methods

A search of MEDLINE database is supplemented by manual searches of bibliographies of key relevant articles. We selected all studies in which a claim code was validated against a medical record. We reported the positive predictive value (PPV) for the VTE claim compared to the medical record.

Results

Our search strategy yielded 345 studies, of which only 19 met our eligibility criteria. All of the studies reported on ICD-9 codes, but only two studies reported on pharmacy codes, and one study reported on procedure codes. The highest PPV (65%–95%) was reported for the combined use of ICD-9 codes 415 (pulmonary embolism), 451, and 453 (deep vein thrombosis) as a VTE event. If a specific event like DVT (PPV 24%–92%) or PE (PPV 31%–97%) was evaluated, the PPV was lower than when the combined events were examined. Studies that included patients after orthopedic surgery reported the highest PPV (96%–100%).

Conclusions

The use of ICD-9 415, 451, and 453 are appropriate for the identification of VTE in claims databases. The codes performed best when codes were evaluated in patients at higher risk of VTE. Copyright © 2012 John Wiley & Sons, Ltd.


INTRODUCTION

  1. Top of page
  2. ABSTRACT
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. ACKNOWLEDGEMENTS
  9. REFERENCES

Venous thromboembolism (VTE) encompasses pulmonary embolism (PE) and deep vein thrombosis (DVT). The incidence of VTE is about 7 per 10,000 person-years among community residents.[1] The condition recurs in about 20% of patients after 5 years, but the rate varies depending on the presence of risk factors.

There are many risk factors for VTE. Common risk factors are surgery, immobilization, and malignancy. Medications are a rare and uncommon risk factor for VTE. Based on an annual incidence of idiopathic VTE in oral contraceptive users of 1–2 cases per 10,000 women per year in the WHO study with a case fatality rate of 1%–2%, the incidence of fatal VTE has been calculated to be 1–4 cases per million women per year.[2]

The Food and Drug Administration Mini-Sentinel contract is a pilot program that aims to conduct active surveillance to detect and refine safety signals that emerge for marketed medical products. To perform this active surveillance, it is necessary to develop and understand the validity of algorithms for identifying health outcomes of interest (HOI) in administrative and claims data. Thus, the goal of this review was to identify algorithms used to detect selected HOI using claims data sources and describe the performance characteristics of these algorithms as reported by the studies in which they were used. This review summarizes the process and findings of the VTE algorithm review.

METHODS

  1. Top of page
  2. ABSTRACT
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. ACKNOWLEDGEMENTS
  9. REFERENCES

Search strategy

Details of the methods for this systematic review can be found in the accompanying manuscript by Carnahan and Moores.[3] In brief, the base PubMed search including terms to identify administrative databases studies with Medical Subject Headings (MeSH) that represented the HOI: “VTE” [MeSH], and “venous thrombosis” [MeSH]. Searches of the citation database of the Iowa Drug Information Service (IDIS/Web) were also conducted. The details of these searches can be found in the full report on the Mini-Sentinel website: http://www.mini-sentinel.org/foundational_activities/related_projects/default.aspx. All searches were conducted in July 2010. All searches were restricted to articles published in 1990 or after. Mini-Sentinel collaborators were also surveyed to help identify any relevant validation studies.

Selection criteria

The abstract of each citation identified was reviewed by two investigators. When either investigator selected an article for full-text review, the full text was reviewed by two investigators. Agreement on whether to review the full text or include the article in the evidence table was calculated using a Cohen's kappa statistic.

Articles were considered for inclusion if they reported on original data where a claim code that identified subjects with VTE was validated against a medical record.

Data abstraction

One investigator was responsible for completing the evidence table (LT), and the second confirmed the accuracy of the data abstracted (TH). Differences between the two reviewers were resolved by consensus. The key exposure variable was the algorithm or code.

The outcome variable of interest was the validation of the algorithm or code against the documentation of VTE in the medical record. Documentation of VTE was defined as a note documenting the occurrence of a VTE or the objective documentation of PE by an angiogram, CT scan or DVT by ultrasound. When possible, the positive predictive value (PPV) of the algorithm or code when compared to the medical record validation was reported.

RESULTS

  1. Top of page
  2. ABSTRACT
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. ACKNOWLEDGEMENTS
  9. REFERENCES

Literature search

Our literature search yielded 345 abstracts, we selected 65 for full-text review; we excluded 37 because they did not study VTE, 193 because they were not administrative and claims database studies, and 50 were excluded because the data source was not from the United States or Canada. Cohen's kappa for agreement between reviewers on inclusion versus exclusion of abstracts was 0.58. The reviewers disagreed mainly on the classification of categories for the excluded articles and not on the inclusion or exclusion of articles.

Of the 65 full-text articles reviewed, ten were included in the final evidence tables; 18 were excluded because the HOI identification algorithm was poorly defined, and 37 were excluded because they included no validation of the outcome definition or reporting of validity statistics. Reviewers identified seven citations for review from full-text article references. Of these, seven were included in the final report. Cohen's kappa for agreement between reviewers on inclusion versus exclusion of full-text articles reviewed was 1.0.

Mini-Sentinel investigators provided no published and no unpublished reports of validation studies that had been completed by their teams.

Of the 19 studies included in the evidence table (Table 1), ten were identified from the initial search strategy, seven were identified from the review of references of studies obtained from the initial strategy, and two were identified from previous research of one of the VTE HOI report authors.[4, 5]

Table 1. Validity of codes for venous thromboembolism
CitationStudy population, time period and inclusion criteriaDescription of outcome studiedAdministrative coding algorithmValidation statistics
Smith et al. 2004[4]Large HMO (Group Health Cooperative) in Washington state included perimenopausal or postmenopausal female aged 30 to 89 years from 1995–2001.Inpatient and outpatient VTE defined as diagnosed with imaging documentationSpecific ICD-9 VTE codes were not reported92% of cases had a positive diagnostic imaging test
Spencer et al. 2009[5]Worcester (MA) Standard Metropolitan Statistical area in 1999, 2001, and 2003.Incident and recurrent VTESpecific ICD-9 VTE codes were not reportedAfter screening more than 7500 potential cases of VTE, a total of 1567 were validated as a first time episode of possible, probable or definite VTE (208 patients had both DVT and PE).
97.3% of the DVT (1310) cases were classified as definite
52.1% of the PE (465) cases were classified as definite
Gertsman et al. 1990[6]Women using oral contraceptives enrolled in Medicaid in the state of Michigan (1980–1986 N = 234,218).Inpatient incident VTE within 56 days of oral contraceptive prescription.ICD-9 codes132 cases of VTE identified with ICD-9 claims. 42% were confirmed as VTE (PPV = 42%).
PE 415.1
VTE 451-453
Probable cases were required to have objective diagnostic test confirmation of disease.PE and VTE ICD-9 codes and outpatient anticoagulant use80 cases of VTE identified using ICD-9 claims and outpatient anticoagulant prescription codes. 65% were confirmed as VTE (PPV = 65%).
Willey et al. 2004[7]Databases from 2 health plans located in the Southeastern and Western United States from 1997 to 2001.Inpatient incident and recurrent DVT and PEVTE ICD-9 codes 451.x and/or 415.1x in either the primary or secondary diagnostic field. More than one outpatient pharmacy claim for anticoagulants (warfarin, heparin, or low-weight molecular heparin).6.7% of charts did not have objective documentation of VTE event (PPV = 93%).
For recurrent VTE events, there was 11% false positive rate (PPV = 89%).
Cushman et al. 2004[8]Cardiovascular health study (1989–1997) and the Atherosclerosis risk in communities study (1987–1996) (n = 756)Incident DVT and PEICD-9 codes:PPV for all codes was 47% and for individual codes were as follows:
Definite DVT was defined as positive duplex or venogram, and PE was defined by ventilation perfusion scan and/or angiogram.PE 451.1xPE 415.1x (n = 153; PPV 72%)
Phlebitis of the deep veins 451.1xPhlebitis of the deep veins 451.1x (n = 74; PPV74%)
Phlebitis of the lower extremities, unspecified 451.2Phlebitis of the lower extremities, unspecified 451.2 (n = 20; PPV 20%)
Phlebitis other sites 451.8Phlebitis other sites 451.8 (n = 15; PPV 13%)
Phlebitis unspecified sites 451.9Phlebitis unspecified sites 451.9 (n = 9; PPV 11%)
Budd Chiari syndrome 453.0Budd Chiari syndrome 453.0 (n = 1; PPV 100%)
Thrombophlebitis migrans 453.1Thrombophlebitis migrans 453.1 (n = 2; PPV 50%)
Thrombosis of the vena cava 453.2Thrombosis of the vena cava 453.2 (n = 6; PPV 50%)
Thrombosis of other specified veins 453.8Thrombosis of other specified veins 453.8 (n = 195; PPV 80%)
Thrombosis of unspecified site 453.9Thrombosis of unspecified site 453.9 (n = 7; PPV 29%)
Placement of vena cava filter 38.7Placement of vena cava filter 38.7(n = 7; PPV 57%)
Henderson et al. 2009[9]Barnes Jewish Hospital (2004). The study included a sample of postsurgical discharge records (n = 3278)Incident inpatient deep vein thrombosis and pulmonary embolism defined as DRGs, anticoagulation codes, and IVC procedure codes.AHRQ Patient Safety Indicator software algorithm. The algorithm includes codes for DVT and PE in the secondary position supplemented with billing and pharmacy natural language processing.PPV 54% (95% CI 45–63)
NPV 99% (95% CI 99–99)
Sensitivity 87%
Specificity 98%
Leibson et al. 2008[10]Mayo Clinic-affiliated hospitals in Minnesota between 1995–1998 (n = 37,845)Objectively confirmed hospital acquired incident venous thromboembolism eventsNeedleman/Buerhaus algorithm plus:Exclusion rule results:
ICD-9 codes:NPV 100%
415.1, 415.11, 451.11, 451.19, 451.2, 451.81, 453.8Specificity 100%
PPV 35%
Sensitivity 74%
Exclusion rule approach (no ICD-9 code 673.2): encounters with one secondary diagnosis of VTEIndicator variable results:
Indicator rule approach: encounters with one secondary diagnosis of VTE claims plus an indicator for hospital acquired condition (not present on admission).NPV 100%
Specificity 100%
PPV 74%
Sensitivity 38%
Kniffin et al. 1994[11]Five percent sample of Medicare enrollees between 1986 and 1989. (n = 7174 PE and n = 8923 DVT).Incidence of PE and DVT from a hospitalization, outpatient, or emergency room visit followed by at least two claims for prothrombin times (PT) during the 2 months after diagnosis.PE: ICD-9 code 415.1 and CPT codes 33910 or 3391591 cases of PE were identified, and 92% of coded cases were confirmed (PPV = 92%).
DVT ICD-9 specific codes: 451.1, 451.11, 451.19, 451.81 453.2.38 cases of DVT were identified using specific codes, and 84% were confirmed (PPV = 84%).
DVT ICD-9 non-specific codes: 451.2, 453.8, 453.9.215 cases of DVT were identified using non-specific codes, and 79% were confirmed as DVT (PPV = 79%).
Spencer et al. 2006[12]Worcester (MA) Statistical Metropolitan Area (12 hospitals) in 1999.Incident recurrent venous thromboembolism classified as definite cases in the outpatient and inpatient setting.ICD-9 codes:2249 subjects were identified as
PE: 415.1(1,9)Of the 142 PE cases, 27% were classified as definite (PPV = 27%).
Confirmed by venography, ultrasonography, CT scan, MRI, or autopsy. Recurrent cases were defined as new occurrence of thrombosis in a previously uninvolved venous or pulmonary segment.DVT: 451, 451.11, 451.19, 451.2, 451.81, 451.83, 451.89, 451.9, 453.1, 453.2, 453.8, 453.9, 671.3 (0,1,3), 671.4 (0,2,4), 671.9 (0 to 4), 673.2(0 to 4), 996.73, 996.74, 997.2Of the 445 DVT cases, 96% were classified as definite (PPV = 96%).
Spencer et al. 2007[13]Worcester (MA) Statistical Metropolitan Area (12 hospitals) in 1999, 2001, and 2003.Incident venous thromboembolism classified as definite cases in the outpatient and inpatient setting if confirmed by venography, ultrasonography, CT scan, MRI, or autopsy.VTE ICD-9 codes:7222 cases were identified using VTE codes, and only 26% of the cases were classified as definite VTE (PPV = 26%).
415.1 (1,9), 451, 451.11, 451.19, 451.2, 451.81, 451.83, 451.89, 451.9, 453.1, 453.2, 453.8, 453.9, 671.3(0,1,3), 671.4(0,2,4), 671.9 (0 to 4), 673.2(0 to 4) 996.73, 996.74, 997.2
Spyropoulos et al. 2002[14]Lovelace Health System serving New Mexico from 1995 to 1998. Included subjects eligible for heparin or warfarin therapy.DVT diagnosed as acute, proximal diagnosed with ultrasound and that oral anticoagulant therapy was planned for at least 3 months.ICD-9 code for DVT: 453.8Of 354 records evaluated, 129 had confirmed DVT (PPV = 34%).
White et al. 2000[15]Medicare inpatient claims of 25,388 fee for service patients who underwent total hip arthroplasty between 1993 and 1996Objective documentation of inpatient incident VTE required positive findings on pulmonary arteriography, V/Q scan, venography, venous ultrasonography, or impedance plethysmography.VTE ICD-9 codes: 451.11, 451.18, 451.2, 451.81, 451.9, 453.1, 453.2, 453.8, 453.9, 415.1Of 297 patients who were hospitalized with postoperative VTE, 285 were objectively confirmed with VTE (PPV 96%).
White et al. 1998[16]California hospital discharge records after unilateral primary total hip (n = 19,586) and knee (n = 24,059) arthroplasty.Incident DVT and PEICD-9 codes for DVT:DVT and PE sensitivity: 100% DVT or PE PPV: 67%
451.11, 451.18, 451.2, 451.81, 451.9, 453.1, 453.2, 453.8, 453.9, 997.2DVT or PE PPV if principal diagnosis at the time of readmission after orthopedic surgery: 100%
ICD-9 codes for PE:DVT or PE PPV for cases treated with an IVC filter: 98%
415.1DVT PPV for cases admitted for 3 or more days with a principal diagnosis of DVT: 92%
McCarthy et al. 2000[17]Medicare hospital discharge data from California and Connecticut for persons 65 and older in 1994 (n = 78)Incident VTE after a medical or surgical hospitalizationICD-9 codes for VTE: 415.1, 451.11, 451.19, 451.2, 451.81, 453.8For postsurgical cases (n = 36), the PPV was 67% when objective clinical evidence of VTE was used as the gold standard, and the PPV was 8% when physician notes were used as the gold standard.
For medical cases (n = 42), the PPV was 55% when objective clinical evidence was used as the gold standard, and the PPV was 12% when physician notes were used.
White et al. 2004[18]UC Davis Medical Center and Kaiser Permanente hospitals female patients before and after childbirth between 1990 and 1998Pregnancy-associated VTEICD-9 VTE codes: 
Each case had to have: (1) objectively confirmed DVT using an objective test; or (2) objectively confirmed PE.400 codes: 451.11, 451.2, 451.8, 451.11, 453.8The PPV and corresponding 95% CI are as follows:
600 codes: 671.30, 671.31, 671.33, 671.42, 671.44, 671.91, 671.92, 671.93, 671.94, 673.21, 673.22, 673.23, 673.24, 673.83400 codes only - 83% (67%–94%)
600 codes only - 30% (23%–37%)
400 or 600 codes as principal diagnosis, hospital stay for 2 days, and a code for a VTE test - 97% (85%–99%)
400 or 600 codes plus VTE test code and hospital stay for more than 3 days - 76% (64%–85%)
400 or 600 codes but excluding 671.31, 671.42, and 671.9x – 65% (56%–74%)
Codes as principal diagnosis alone – 73% (61%–82%)
Codes as principal diagnosis plus stay for 2 days or more - 79% (67%–87%)
Codes as principal diagnosis plus stay for 2 days or more and VTE test code – 97% (85%–99%)
415.1x, 451.1 453.8, 671.33, and 673.2 -77% (66%–86%)
Codes as principal diagnosis alone - 87% (73%–95%)
Codes as principal diagnosis plus stay for 2 days or more - 95% (83%–99%)
Codes as principal diagnosis plus stay for 2 days or more and VTE test code - 100% (85%–100%)
Zhan, C. et al. 2007[26]The Medicare Patient Safety Monitoring System 2002–2004 (CMS). N = 20,868 eligible surgical hospitalizations.Confirmed postsurgical DVT or PEICD-9 codes for DVT:Of 20,868 eligible hospitalizations, 232 DVT cases and 95 PE cases were identified by ICD-9 codes, 72 of the DVT cases and 23 of the PE cases confirmed by medical chart abstraction:
451.11, 451.19, 451.2, 451.81, 451.9, 453.40, 453.41, 453.42, 453.8, 453.9
ICD-9 codes for PE:
415.1, 451.11, 451.19
 DVT PPV 31%
PE PPV 24%
PE/DVT combined PPV 29%
DVT sensitivity 67%
PE sensitivity 74%
PE/DVT sensitivity 68%
Anderson et al. 1991[27]Worcester (MA) Standard Metropolitan Statistical area population discharged with a diagnosis of DVT or PE from 1985–1986 (n = 405).Incident initial hospital discharge of VTEICD-9 codes for VTE: 415.1, 451.11, 451.19, 451.2, 451.81, 453.2, 453.8, 453.9, 639.6, 671.30, 671.31, 671.33, 671.40, 671.42, 671.44, 671.91, 997.2, 999.2, 996.7230 of 274 patients had objectively confirmed DVT (PPV = 84%)
DVT was confirmed by venography, impedance plethysmosgraphy, or ultrasound and PE by pulmonary angiography and/or lung scan.71 of 131 patients had objectively confirmed PE. (PPV was 54%)
Weingart et al. 2000[28]Medicare hospital discharge data from California and Connecticut for persons 65 and older in 1994 (n = 1025)Incident VTE after a medical or surgical hospitalizationICD-9 codes for VTE 415.1, 451.11, 451.19, 451.2, 451.81, 453.8For patients with a suspected in-hospital VTE (n = 40), the PPV was 70%.
For patients with a suspected medical case of VTE (n = 71), the PPV was 28%.
Lawthers et al. 2000[29]Medicare hospital discharge data from California and Connecticut in 1994.Incident VTE after medical or surgical hospitalizationICD-9 codes for VTE 415.1, 451.11, 451.19, 451.2, 451.81, 453.8For surgical cases (n = 813):
PPV 90%
NPV 98%
For medical cases (n = 485):
PPV 76%
NPV 99%

The possible reasons why the two articles were missed by the search strategy were the fact that the purpose of the articles was to evaluate the risk and trends of VTE in specific population and not the validity of the codes.

The purpose of all but two studies[4, 6] was to evaluate the incidence of VTE in postsurgical or medical hospital admissions and to validate the codes utilized to identify cases in claims databases against medical records.

Baseline characteristics of the included studies

Of the 19 studies, only one evaluated claims databases from a national private health insurance company.[7] Ten studies evaluated private state/local health plans, and seven evaluated Medicare and Medicaid databases. One study used a cohort of subjects in the Cardiovascular Health Study and Atherosclerosis Risk in Communities Study.[8]

Fifteen of the 19 studies listed in the evidence table reported the algorithms of International Classification of Diseases, Ninth Revision (ICD-9), codes to identify patients with VTE, the remaining four studies[4, 5, 9, 10] did not report the codes necessary for replication but mentioned that ICD-9 codes were used.

Two studies added pharmacy codes to ICD-9 codes,[6, 7] and one study added CPT codes to ICD-9 codes[9] when validating against medical records. No study compared ICD-10 or DRG codes with medical records for the validation of VTE. All studies included in the report validated administrative and claims data algorithms through abstraction of medical charts. Documentation of VTE in the medical records was generally based on clinical encounters. Fifteen studies evaluated codes for the occurrence of VTE in an inpatient setting or as a complication of medical or surgical inpatient stay. Four studies included study populations from inpatient, outpatient, or emergency room settings.[4, 11-13]

Ten studies used confirmation of the diagnosis by imaging, and only one study[17] reported on both methods of validation. For postsurgical patients with claims for VTE, the PPV was 67% when objective evidence of VTE compared to a PPV of 8% when physician notes were used. The same trend was seen in medical patients. Three studies used validation criteria classifying patients by categories of suspicion of VTE as definite, probable, and negative.[6, 12, 13]

For the purpose of this review, we used the definite category for the diagnosis of VTE. All other studies classified subjects as having or not having VTE.

Validation algorithms

Five studies evaluated several codes for PE. The range of PPV for the PE codes was 24% to 92%. Two studies used a single ICD-9 code – 415.1 (PE and infarction) – and reported a PPV that ranged from 27% to 72%.[8, 12] A potential explanation for the low performance of ICD-9 code 415.1 in the study by Spencer et al. was that subjects included were recurrent cases of PE. One study used the 415.1 code for PE, CPT codes 33910 and 33915 (pulmonary embolectomy), and at least two claims for prothrombin time and reported a PPV of 92%.[11] The study by Zhan et al. reported a PPV of 24%; however, to identify PE patients, the study used the 415.1 code but also DVT codes (451.1).

Nine studies evaluated several codes for DVT. The range of PPV for the DVT codes was 31% to 97%. One study used a single ICD-9 code – 453.8 (other venous embolism and thrombosis) – and reported a PPV of 36%.[14] One study[8] compared the PPV for single ICD-9 codes for the diagnosis of DVT and reported a PPV of 80% for 453.8, 74% for 451.1 (phlebitis and thrombophlebitis), and 50% for 453.2 (vena cava).

When ICD-9 codes 451 and 453 (of the vena cava) as well as the subcodes that indicate the vein where the thrombosis occurred were used, the reported PPV was 84%. For non-specific codes 451.2, 453.8, and 453.9, the reported PPV was 79%.[11]

For VTE defined as either PE or DVT, several codes were evaluated. When ICD-9 codes and subcodes 451.x, 453.x, and 415.x were used, the reported PPV was 96%.[15] When ICD-9 codes and subcodes 451.x and 415.x were used, the reported PPV was 93%.[18]

Two studies reported on the validity of VTE codes combined with pharmacy claims in their algorithms. The first[7] reported a PPV of 93% when VTE codes 415.1 and 451.x were used with at least one outpatient pharmacy claim for anticoagulants. The second[6] reported a PPV of 42% when the VTE codes 451–453 were used; however, when the same codes were added to outpatient anticoagulant claim, the PPV increased to 65%.

Specific populations

Two studies evaluated the PPV for VTE codes in the setting of hip and knee surgery. The first, reported by White et al., found that VTE codes (451.x, 453.x, 415.1) in Medicare patients who underwent total hip arthroplasty had a PPV of 96%. The second[15] found that VTE codes (451.x, 453.x, 997.2, 415.1) in hospital discharge records of patients who underwent unilateral total primary hip and knee arthroplasty had a PPV of 100% at the time of readmission.

One study evaluated the validity of VTE codes in women using oral contraceptives enrolled in Medicaid and found a PPV of 42%. When codes 671 (pregnancy related VTE), 996, and 997.3 (phlebitis or embolism due to presence of device) were added to codes 415.x, 451.x, 453.x, the reported PPV was 67%. One study evaluated the validity of obstetrical codes (600 codes); when used alone, the reported PPV was 30%. When codes 671.x, 673.x were added to codes 451.x and 453.x, the reported PPV was 71%.[18]

Three studies compared the performance of VTE codes (415.x, 451.x, 453.8) in a Medicare population in both medical and postsurgical patients. For postsurgical patients the PPV of the previously mentioned ICD-9 codes ranged from 67% to 90% compared to 55% to 76% for medical patients.

DISCUSSION

  1. Top of page
  2. ABSTRACT
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. ACKNOWLEDGEMENTS
  9. REFERENCES

This systematic review found that the combined use of ICD-9 codes 415.x (PE), 451.x, and 453.x (DVT) as a VTE event consistently yielded the highest PPV.

The study also found that if a specific event like DVT or PE was evaluated, the PPV was lower than when the combined events were examined. The range of PPVs was similar with codes for PE when compared to codes for DVT.

The performance of the codes when compared to medical records was dependent on the population evaluated. Codes evaluated after orthopedic procedures or in surgical patients had the highest PPVs; however, studies reporting codes in women using oral contraceptives or after delivery reported lower PPVs.

Our systematic review has several limitations that deserve mention. First, and probably the most important limitation, is the lack of standard reporting of the objective criteria used for the diagnosis of VTE. This is of particular importance when considering that the diagnostic accuracy is different if a CT scan or a V/Q is used for the diagnosis of PE. Second, there was a paucity in reporting the use of other claims data like pharmacy, procedure, and DRG codes. Third, the influence of the prevalence on the performance of the PPV in high-risk patients (orthopedic patients) has not been adequately explored.

The performance of VTE ICD-9 codes was dependent on the population. ICD-9 codes evaluated in higher risk subjects like postsurgical patients or Medicare patients consistently reported the highest PPV. This finding is consistent with the fact that the incidence among average-risk medical patients is 10% to 20% and rises to 80% in the spinal cord injury, major trauma, and critically ill patient populations.[19] VTE is idiopathic 25% to 50% of the time.[20] However, this was not properly addressed by all the studies included in our systematic review.

VTE often results from the interactions between several risk factors. In a review of 1231 consecutive patients treated for VTE, 96% had > 1 recognized risk factor.[21] Medications have been recognized as a potential risk factor for VTE in both men and women. Case-control studies concluded that the incidence of VTE in young and otherwise healthy women taking oral estrogen contraceptives is between 1 and 3 per 10 000 per year.[22]

Also, women using hormone replacement therapy have a 2- to 4-fold increased risk of idiopathic venous thrombosis compared with women not receiving HRT.[2] In the Heart and Estrogen/Progestin Replacement Study, the relative hazard ratio for VTE in postmenopausal women taking estrogen plus progestin was 2.7.[23] Men receiving estrogen therapy for prostate cancer are also at increased risk for VTE.[24]

Our study also found that the use of ICD-9 codes that indicate VTE in pregnant women (ICD-9 codes in the 600 range) yielded lower PPV than codes normally used for all patients (ICD-9 codes in the 400 range). This is particularly relevant because pregnancy increases this risk five times. PE is a leading cause of maternal death after childbirth, with one clinically recognized PE per 1000 births and one fatal PE per 100 000 births.[25]

In conclusion, there is a need for active surveillance of potentially fatal complications in high-risk patients, and this study provides a set of codes that can be used for active surveillance. We recommend the combined use of ICD-9 codes 415.x, 451.x, and 453.x particularly in high-risk patients (orthopedic surgery or surgical admissions) to identify patients with VTE. We also recommend that future work be conducted to determine the sensitivity of claims data for identifying cases of VTE, using other claims and administrative dataset elements in addition to ICD-9 codes.

CONFLICT OF INTEREST

  1. Top of page
  2. ABSTRACT
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. ACKNOWLEDGEMENTS
  9. REFERENCES

The authors have declared that there is no conflict of interest.

KEY POINTS

  • ICD-9 codes can identify VTE events in claims databases
  • The most commonly used databases were Medicare and Medicaid
  • The use of ICD-9 415, 451, and 453 is appropriate for the identification of VTE.

ACKNOWLEDGEMENTS

  1. Top of page
  2. ABSTRACT
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. ACKNOWLEDGEMENTS
  9. REFERENCES

The views expressed in this document do not necessarily reflect the official policies of the Department of Health and Human Services (HHS), nor does mention of trade names, commercial practices, or organizations imply endorsement by the U.S. government. This work was supported by the Food and Drug Administration through Department of HHS Contract Number HHSF223200910006I.

REFERENCES

  1. Top of page
  2. ABSTRACT
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. ACKNOWLEDGEMENTS
  9. REFERENCES
  • 1
    Heit JA, Melton LJ, III, Lohse CM, et al. Incidence of venous thromboembolism in hospitalized patients vs community residents. Mayo Clin Proc 2001; 76(11): 11021110.
  • 2
    Venous thromboembolic disease and combined oral contraceptives: Results of international multicentre case-control study. world health organization collaborative study of cardiovascular disease and steroid hormone contraception. Lancet 1995; 346(8990): 15751582.
  • 3
    Carnahan RM, Moores KG. Mini-Sentinel's systematic reviews of validated methods for identifying health outcomes using administrative data: methods and lessons learned. Pharmacoepidemiol Drug Saf 2012; 21(S1): 8289.
  • 4
    Smith NL, Heckbert SR, Lemaitre RN, et al. Esterified estrogens and conjugated equine estrogens and the risk of venous thrombosis. JAMA 2004; 292(13): 15811587.
  • 5
    Spencer FA, Emery C, Joffe SW, et al. Incidence rates, clinical profile, and outcomes of patients with venous thromboembolism. the worcester VTE study. J Thromb Thrombolysis 2009; 28(4): 401409.
  • 6
    Gerstman BB, Freiman JP, Hine LK. Use of subsequent anticoagulants to increase the predictive value of medicaid deep venous thromboembolism diagnoses. Epidemiology 1990; 1(2): 122127.
  • 7
    Willey VJ, Bullano MF, Hauch O, et al. Management patterns and outcomes of patients with venous thromboembolism in the usual community practice setting. Clin Ther 2004; 26(7): 11491159.
  • 8
    Cushman M, Tsai AW, White RH, et al. Deep vein thrombosis and PE in two cohorts: The longitudinal investigation of thromboembolism etiology. Am J Med 2004; 117(1): 1925.
  • 9
    Henderson KE, Recktenwald A, Reichley RM, et al. Clinical validation of the AHRQ postoperative venous thromboembolism patient safety indicator. Jt Comm J Qual Patient Saf 2009; 35(7): 370376.
  • 10
    Leibson CL, Needleman J, Buerhaus P, et al. Identifying in-hospital venous thromboembolism (VTE): A comparison of claims-based approaches with the rochester epidemiology project VTE cohort. Med Care 2008; 46(2): 127132.
  • 11
    Kniffin WD, Jr, Baron JA, Barrett J, Birkmeyer JD, Anderson FA, Jr. The epidemiology of diagnosed pulmonary embolism and deep venous thrombosis in the elderly. Arch Intern Med 1994; 154(8): 861866.
  • 12
    Spencer FA, Emery C, Lessard D, et al. The worcester venous thromboembolism study: A population-based study of the clinical epidemiology of venous thromboembolism. J Gen Intern Med 2006; 21(7): 722727.
  • 13
    Spencer FA, Lessard D, Emery C, Reed G, Goldberg RJ. Venous thromboembolism in the outpatient setting. Arch Intern Med 2007; 167(14): 14711475.
  • 14
    Spyropoulos AC, Hurley JS, Ciesla GN, de Lissovoy G. Management of acute proximal deep vein thrombosis: Pharmacoeconomic evaluation of outpatient treatment with enoxaparin vs inpatient treatment with unfractionated heparin. Chest 2002; 122(1): 108114.
  • 15
    White RH, Gettner S, Newman JM, Trauner KB, Romano PS. Predictors of rehospitalization for symptomatic venous thromboembolism after total hip arthroplasty. N Engl J Med 2000; 343(24): 17581764.
  • 16
    White RH, Romano PS, Zhou H, Rodrigo J, Bargar W. Incidence and time course of thromboembolic outcomes following total hip or knee arthroplasty. Arch Intern Med 1998; 158(14): 15251531.
  • 17
    McCarthy EP, Iezzoni LI, Davis RB, et al. Does clinical evidence support ICD-9-CM diagnosis coding of complications? Med Care 2000; 38(8): 868876.
  • 18
    White RH, Brickner LA, Scannell KA. ICD-9-CM codes poorly indentified venous thromboembolism during pregnancy. J Clin Epidemiol 2004; 57(9): 985988.
  • 19
    Geerts WH, Bergqvist D, Pineo GF, et al. Prevention of venous thromboembolism: American college of chest physicians evidence-based clinical practice guidelines (8th edition). Chest 2008; 133(6 Suppl): 381S453S.
  • 20
    Blann AD, Lip GY. Venous thromboembolism. BMJ 2006; 332(7535): 215219.
  • 21
    Anderson FA, Jr, Wheeler HB. Physician practices in the management of venous thromboembolism: A community-wide survey. J Vasc Surg 1992; 16(5): 707714.
  • 22
    Hedenmalm K, Samuelsson E. Fatal venous thromboembolism associated with different combined oral contraceptives: A study of incidences and potential biases in spontaneous reporting. Drug Saf 2005; 28(10): 907916.
  • 23
    Hulley S, Grady D, Bush T, et al. Randomized trial of estrogen plus progestin for secondary prevention of coronary heart disease in postmenopausal women. heart and Estrogen/progestin replacement study (HERS) research group. JAMA 1998; 280(7): 605613.
  • 24
    Lundgren R, Sundin T, Colleen S, et al. Cardiovascular complications of estrogen therapy for nondisseminated prostatic carcinoma. A preliminary report from a randomized multicenter study. Scand J Urol Nephrol 1986; 20(2): 101105.
  • 25
    Hogberg U. Maternal deaths in sweden, 1971–1980. Acta Obstet Gynecol Scand 1986; 65(2): 161167.
  • 26
    Zhan C, Battles J, Chiang YP, Hunt D. The validity of ICD-9-CM codes in identifying postoperative deep vein thrombosis and pulmonary embolism. Jt Comm J Qual Patient Saf 2007; 33(6): 326331.
  • 27
    Anderson FA, Jr, Wheeler HB, Goldberg RJ, et al. A population-based perspective of the hospital incidence and case-fatality rates of deep vein thrombosis and pulmonary embolism. the worcester DVT study. Arch Intern Med 1991; 151(5): 933938.
  • 28
    Weingart SN, Iezzoni LI, Davis RB, et al. Use of administrative data to find substandard care: Validation of the complications screening program. Med Care 2000; 38(8): 796806.
  • 29
    Lawthers AG, McCarthy EP, Davis RB, Peterson LE, Palmer RH, Iezzoni LI. Identification of in-hospital complications from claims data. is it valid? Med Care 2000; 38(8): 785795.