DVT Paper

O R I G I N A L A R T I C L E

STOPDVTs: Development and testing of a clinical assessment tool to guide nursing assessment of postoperative patients for Deep Vein Thrombosis

Alanna O’Brien BN, MN, Director Clinical Nursing Education1 | Bernice Redley PhD, BN,

Associate Professor Nursing Research2,3 | Beverley Wood PhD, BN, Research

Fellow1,2 | Mari Botti PhD, BN, Chair in Nursing1,2 | Anastasia F Hutchinson PhD, BN,

Associate Professor of Nursing Research1,2

1Centre for Quality and Patient Safety

Research-Epworth HealthCare Partnership,

Deakin University, Geelong, Vic., Australia

2School of Nursing and Midwifery, Deakin

University, Geelong, Vic., Australia

3Centre for Quality and Patient Safety

Research-Monash Health Partnership,

Monash Health, Deakin University, Clayton,

Vic., Australia

Correspondence

Anastasia F. Hutchinson, Anastasia School

of Nursing and Midwifery and Centre for

Quality and Patient Safety Research

Epworth HealthCare partnership, Geelong,

Vic., Australia.

Email: [email protected]

Aims and objectives: To develop and test a clinical tool to guide nurses’ assessment

of postoperative patients for Deep Vein Thrombosis.

Background: Preventing venous thromboembolism in hospitalised patients is an

international patient safety priority. Despite high-level evidence for optimal venous

thromboembolism prophylaxis, implementation is inconsistent and the incidence of

Deep Vein Thrombosis remains high.

Methods: A two-stage sequential multi-method design was used. In stage 1, the

STOPDVTs tool was developed using a review of the literature and focus groups

with local clinical experts. Stage 2 involved pilot testing the tool with 38 surgical

nurses who conducted repeated assessments on a prospective sample of 50 postop-

erative orthopaedic patients.

Results: Stage 1: The focus group members who were members of the nursing lead-

ership team agreed on eight local and systemic signs and symptoms that should be

included in a nursing patient assessment tool for early Deep Vein Thrombosis. Local

symptoms were pain in the limbs, calf swelling and tightness, changes in the affected

limb’s skin temperature. Systemic signs included in the tool were as follows: increased

shortness of breath, increased respiratory and heart rates, and decreased oxygen satu-

ration. Stage 2: The STOPDVTs tool had acceptable face and content validity, the

agreement between the expert nurse and surgical nurses on assessments of individual

signs and symptoms varied between 44%–94%. Surgical nurses were less likely than

the expert nurse to identify signs indicative of Deep Vein Thrombosis.

Conclusion: Despite finding the STOPDVTs clinical assessment tool was a useful

guide for nursing assessment, surgical nurses often underestimated the potential

importance of clinical signs. The findings reveal a gap in nursing knowledge and skill

in assessing for Deep Vein Thrombosis in postoperative orthopaedic patients.

Relevance to clinical practice: This study identified a possible risk to patient safety

related to under-recognition of the signs and symptoms of possible Deep Vein

Thrombosis (DVT) in postoperative orthopaedic patients. The findings demonstrate

Accepted: 18 February 2018

DOI: 10.1111/jocn.14329

J Clin Nurs. 2018;27:1803–1811. wileyonlinelibrary.com/journal/jocn © 2018 John Wiley & Sons Ltd | 1803

the feasibility of developing and implementing a protocol for consistent screening

by nurses for possible DVT in the postoperative period.

K E Y W O R D S

clinical assessment, clinical assessment tools, deep vein thrombosis, orthopaedic nursing,

surgical nursing, venous thromboembolism

1 | BACKGROUND

The prevention and management of venous thromboembolism (VTE)

are recognised internationally as a major patient safety priority (Gat-

son, White, & Misan, 2012; Australian National Safety and Quality

Health Service Standards (ACSQHC, 2011); National Health and

Medical Research Council (NHMRC), 2009). Despite significant

improvements in VTE management over the last decade, there are

still many potentially preventable deaths due to missed or late diag-

nosis (Bacon, 2013). In Australia, in 2008, there were approximately

2,000 deaths attributed to VTE that may have been avoided if

patients’ future VTE risk had been assessed using a specific clinical

risk assessment framework and appropriate prophylaxis commenced

(NHMRC, 2009).

The prevalence of Deep Vein Thrombosis (DVT) following hip or

knee surgery is amongst the highest of all surgical specialties (Baser,

Supina, Sengupta, Wang, & Kwong, 2010; Falck-Ytter et al., 2012;

Maletis, Inacio, Reynolds, & Funahashi, 2012). DVT is common after

orthopaedic surgery due to the damage caused to the endothelial

layer of the tissues which triggers heightened coagulation, leading to

platelet activation and clot formation (Emadi & Streiff, 2011). DVT is

the most common cause for readmission after hip surgery and is a

major preventable cause of postoperative complications in the

orthopaedic joint replacement surgery population (Bottaro et al.,

2012; Falck-Ytter et al., 2012; Kanchanabat et al., 2011). Clinical tri-

als have shown that 15%–20% of patients who have undergone

major joint replacement surgery may leave hospital with asymp-

tomatic DVT (detectable only on venography ultrasound), despite

being commenced on thromboprophylaxis therapy (Dahl, Gudmund-

sen, & Haukeland, 2000).

Despite its long history in clinical medicine, the diagnosis of DVT

remains problematic, clinical signs are nonspecific meaning that

potential cases of early DVT may go undetected by clinical staff (Ali

& Young, 2012; Bacon, 2013; Songwathana, Promlek., & Naka,

2011). The typical signs and symptoms of DVT include pain, swelling,

tenderness, muscle cramps, discoloration or redness of the affected

area and skin that is warm to touch (Bauersachs, 2012; Tabei, Holtz,

Sch€urer-Maly, & Abholz, 2012). Previous studies have shown that

clinical features alone have limited diagnostic value in DVT diagnosis

with the highest reported positive predictive value being approxi-

mately 60% (Goodacre, Sutton, & Sampson, 2005; Wells & Ginsberg,

1995). Homan’s sign (pain in the calf when there is dorsiflexion of

the foot) is no longer routinely used as a diagnostic tool for DVT

due to its lack of sensitivity and specificity (Bacon, 2013). The rea-

sons for the poor prognostic value of clinical signs are that each

patient may present differently depending on site and size of throm-

bus and that the clotting process can also mimic many other disease

states (Hotoleanu, Fodor, & Suciu, 2010; Strijkers, Cate-Hoek, Buk-

kems, & Wittens, 2011; Tan, Van Rooden, Westerbeek, & Huisman,

2009; Tenna, Kappadath, & Stansby, 2012). Assessing postoperative

orthopaedic patients also poses its own unique challenges, particu-

larly as the surgical procedure also contributes to swelling and

oedema in the affected limb (Schiff et al., 2005).

Nurses play a critical role in the detection of early DVT in

postoperative patients as they conduct frequent patient assess-

ments, follow patients through their inpatient journey and have the

ability to prompt early investigation of potential cases and escala-

tion of care (Schiff et al., 2005; NHMRC, 2011; Tabei et al., 2012).

Despite its importance, guidance to assist nurses with specific clini-

cal assessments to detect possible DVT in the orthopaedic postop-

erative setting is absent.

Due to these limitations, there is currently no identified “gold

standard” in clinical assessment for the detection of early DVT.

Despite these well-known challenges, there is a need for clinicians,

particularly nurses, to have a high level of awareness of the subtle

changes associated with possible DVT, so that potentially affected

patients can be identified early and care escalated to provide further

diagnostic tests (such as ultrasound), so that a definitive diagnosis

can be made and treatment instigated. There are currently no

What does this paper contribute to the wider

global clinical community?

• This study identified a possible risk to patient safety

related to under-recognition of the signs and symptoms

of possible Deep Vein Thrombosis (DVT) in postoperative

orthopaedic patients.

• The findings demonstrate the feasibility of developing

and implementing a protocol for consistent screening by

nurses for possible DVT in the postoperative period.

• This study highlights the need for further translational

research to be conducted to embed both risk screening

for venous thromboembolism and clinical assessment for

DVT into clinical practice.

1804 | O’BRIEN ET AL.

published tools available to assist nurses to systematically assess

patients for the range of possible signs and symptoms of early DVT

in the postoperative setting (Righini & Bounameaux, 2007). The pur-

pose of this study was to provide a valid, reliable and usable clinical

assessment tool to guide nurses’ assessment of postoperative

patients for DVT.

1.1 | Aims

The study aims were to:

1. develop a DVT clinical assessment tool to assist nurses conduct a

standardised, systematic and comprehensive assessment and

enhance documentation of signs and symptoms of early DVT in

postoperative patients and,

2. assess (i) the face validity and usability of the tool in clinical prac-

tice and (ii) to measure the level of inter-rater agreement

between assessors using a sample of surgical nurses and high-risk

postoperative orthopaedic patients.

2 | METHODS

2.1 | Research design

This research project used a two-stage sequential multi-method

study design: the first stage of the project used a qualitative

methodology to support the development of the STOPDVTS clinical

assessment tool. In the second stage, an observational quantitative

methodology was used to evaluate a pilot implementation of the

clinical assessment tool into clinical care at a single site.

2.1.1 | Study setting

The study was conducted on two acute orthopaedic surgery wards

in a large, private healthcare service in Melbourne, Australia.

2.2 | Stage 1. Development of the STOPDVTS clinical assessment tool

In Stage 1, the new tool was developed. The components of the tool

were derived from analysis of relevant literature and evaluation of

face validity used feedback from a panel of expert nurses. This stage

of the research involved two steps.

2.2.1 | Review of the research evidence

A systematic search of the research literature between 1995–2015

was undertaken to identify VTE risk assessment scores and DVT

clinical assessment tools. The search terms used included; “Venous

Thromboembolism,” “Venous thrombosis,” “Venous thrombosis pre-

vention and control,” “Deep Vein Thrombosis,” “diagnosis,” “assess-

ment,” “evidence based practice,” “orthopaedic surgery,” “nursing

interventions,” “nursing assessment,” “nursing process,” “nursing

role,” “nursing skills,” “patient assessment,” “risk assessment,” and

clinical assessment tools were used to gather relevant research arti-

cles. A total of 45 relevant articles were identified.

The Wells score was identified as the most widely studied clini-

cal prediction tool for the identification of DVT (Dewar & Corretge,

2007; Penaloza, Laureys, Wautrecht, Lheureux, & Motte, 2006;

Wells & Anderson, 2013; Wells & Ginsberg, 1995; Wells & Scarvelis,

2006; Wells et al., 2010), hence formed the basis for the tool

FIGURE 1 Differences between the expert nurse and the surgical nurses in identification of signs and symptoms of Deep Vein Thrombosis (DVT) [Colour figure can be viewed at wileyonlinelibrary.com]

O’BRIEN ET AL. | 1805

development. The criteria from the Wells score were reviewed to

determine ease of application by nurses at the bed-side and appro-

priate guidance was incorporated into the new clinical assessment

tool. The new tool (known by mnemonic “STOPDVTS”) was designed

to prompt nurses about what to look for when assessing postopera-

tive patients at risk for the development of a possible DVT (Fig-

ure 1).

The STOPDVTS tool was designed to complement the VTE risk

assessment tool that is performed on admission to acute care and

includes background risk factors such as age, history of DVT, pre-

existing clotting disorders and patients undergoing major orthopaedic

surgery (Autar, 2003; Maynard et al., 2010).

2.2.2 | Focus group with nurse experts

Using a focus group, nurses who were members of the nursing lead-

ership team (nurse unit managers, clinical nurse educators) were

invited to attend a focus group discussion to evaluate the draft

STOPDVTS tool in terms of the content, comprehensiveness, tool

layout and ease of use. The panel of eight nursing leaders that rep-

resented all the specialty areas of the hospital expected to assess

patients for early DVT (intensive care, cardiac, vascular, medical sur-

gical, orthopaedics and emergency) were recruited and gave written

informed consent to focus group participation. During the focus

group, the content and design of the draft STOPDVTS clinical assess-

ment tool were discussed. Following the focus group discussion, the

draft STOPDVTS tool was edited in response to feedback from the

participants and then re-circulated electronically to all members of

the focus group panel for comments and approval. This process was

repeated four times until 100% agreement was reached (Table 1).

2.3 | Stage 2. Pilot implementation of the STOPDVTS tool

In Stage 2, a pilot implementation study was undertaken on acute

two orthopaedic wards.

Two participant samples were used in this stage of study; nurse

participants and patient participants.

2.3.1 | Nurse participants

A convenience sample of 38 surgical nurses from two orthopaedic

surgery wards was recruited. All surgical nurse participants attended

training in how to use the STOPDVTS tool (see implementation

below). The nurse caring for an eligible patient was invited to pro-

vide verbal consent to participate on the day of data collection.

2.3.2 | Patient participants

A convenience sample of 50 consecutive adult postoperative

patients who had either total knee (total knee replacement [TKR]) or

hip (total hip replacement [THR]) joint replacement surgery was

recruited over two months. Exclusion criteria for patient participants

were those who declined or were unable to give verbal consent to

study participation. On postoperative day 1, the nurse in charge was

consulted to identify eligible patients on the inpatient ward. The

nurses caring for these patients were asked for an appropriate time

to approach the patients to recruit and obtain verbal consent. The

sample size and repeated assessments (up to three) were expected

to reflect expected usual practice and provide sufficient variability in

normal and abnormal assessment findings to examine the usability of

the STOPDVTS clinical assessment tool.

2.4 | Study procedures

2.4.1 | Implementation of the tool

Implementation was informed by an evidence-based knowledge transla-

tion model (Kitson et al., 2008) that used the following strategies to

implement the STOPDVTS tool on the participating wards. Existing qual-

ity improvement work occurring on the wards to improve VTE manage-

ment was used to support implementation of the new tool; these

included improvements in VTE prophylaxis practices, clinical handover,

VTE risk assessment and systematic patient assessment. Similarly, edu-

cation sessions were conducted for clinical nursing staff and the DVT

screening and assessment processes were integrated into existing risk

screening strategies (Grimshaw, Eccles, Lavis, Hill, & Squires, 2012).

Over four education sessions, 38 surgical nurses were trained in

how to use the tool: sessions included education about: risks for and

prevalence of DVT after orthopaedic surgery; common signs and

symptoms for DVT, discussion of the mnemonic “STOPDVTS” and

the rationale for including each of the signs and symptoms in the

assessment tool. The aim was to increase surgical nurses’ awareness

of their important role in early DVT detection and escalation of care.

Education was reinforced with posters displayed in the ward areas,

emails sent from the nurse managers to staff, lanyard cards and

instruction sheets given to all nursing staff (Grimshaw et al., 2012).

2.5 | Data collection

Stage 2 data were collected from a prospective cohort of 50 inpa-

tients following orthopaedic joint replacement surgery. Each patient

TABLE 1 STOPDVTS Mnemonic

Assess the patient for the presence of the following eight signs and

symptoms:

S—Swelling and Shortness of breath

T—Skin on the affected limb that is hot or cold to Touch

—The presence of Tachycardia

O—Operation (Has the patient undergone a total hip or knee

replacement?)

P—Pain (Has the patient got increased pain in the affected limb?)

D—Discoloration (Is the skin on the affected limb discoloured?)

V—Veins/Varicoses (Does the patient have swollen/distended

varicose veins in the affected limb?)

T—Time (How many days postsurgery)

S—Still/Sedentary (The patient is more immobile than expected)

1806 | O’BRIEN ET AL.

was assessed for DVT using the STOPDVTS tool on up to three

occasions: on postoperative Days 1, 3 and 5 (or day of discharge if

earlier). Data collection from patients involved two independent

physical examinations: one by the expert nurse and the other by a

surgical nurse working that day.

The nurse expert (who provided the “reference standard”) was a

critical care trained nurse with over ten years of experience. She was

also trained specifically in DVT assessment for this study by a surgeon

who also had a specialty interest in VTE prophylaxis and management.

This training had involved demonstration of the steps involved in con-

ducting a clinical assessment for potential DVT and performing super-

vised clinical assessments on a series of inpatients with and without

actual DVTs to demonstrate their assessment skills.

The surgical nurse participants were invited to examine the

patient participants using the STOPDVTS tool, record their findings

on a specific data collection tool developed for the study and place

their findings in a sealed envelope. For each patient, the two exami-

nations occurred independently, but within the same shift. The qual-

ity of the surgical nurses’ documentation of their patient assessment

in the patient medical record was also audited using a purpose-speci-

fic tool. Patients’ medical records were also reviewed to determine

the incidence of new onset DVTs over 21 days postsurgery. All audit

data were collected by the nurse expert.

2.6 | Data analysis

Quantitative data collected during the patient assessments were coded,

collated in a database and analysed using descriptive statistics. Differ-

ences between patient groups from the two participating wards were

analysed using chi-square tests for categorical variables and unpaired t

tests for continuous variables. To evaluate usability, the percentage

agreement between the nurse expert and the surgical nurse assess-

ments of the same patient with the STOPDVTS clinical assessment tool

was compared. The level of agreement between the ward nurse and the

expert nurse was summarised as percentage agreement. The quality of

the ward nurses’ documentation of their patient assessment in the

patient medical record was also used to examine usability of the new

tool, as the tool provided instruction for documentation in care records.

Ethics approval was obtained from the institutional Human

Research Ethics Committee HREC EP174-14.

3 | RESULTS

3.1 | Stage 1: Focus groups

Analysis of transcripts from the focus group with the eight expert

nurses was used to examine the face and content validity, and usability

of the components included in the STOPDVTS clinical assessment tool.

3.1.1 | Content validity of the DVT assessment tool

The key signs and symptoms identified by the nurse experts were

consistent with those identified in the literature, hence, supported

tool content validity. The experts agreed on signs and symptoms

that should be included in standard patient assessment for early

DVT: pain in the limbs, calf swelling and/tightness and changes in

limb skin temperature, discoloration of the affected limb, and the

presence of distended veins or varicoses. In addition, they agreed

patients should be assessed for the presence of the following sys-

temic symptoms: increased shortness of breath, increased respiratory

rate, decreased oxygen saturations and tachycardia, as the presence

of one or more of these symptoms may suggest possible pulmonary

embolus.

3.2 | Stage 2: Pilot implementation study

The implementation study was conducted on two acute orthopaedic

surgery wards (Ward A and Ward B). Fifty prospectively recruited

patients were assessed over days 1 to 5 postoperatively using the

STOPDVTs clinical assessment tool, and a total 114 paired assess-

ments were performed.

There were 38 (five male and 33 female) surgical nurses who

participated as DVT assessors in the project, their mean age was 32

(range 21–71) years. Eighty per cent of nurse participants were

experienced orthopaedic surgery nurses and 20% were new graduate

nurses with 1 to 2 years of clinical experience. Eleven (29%) had

postgraduate qualifications in advanced nursing and were working as

clinical nurse specialists, or in associate charge nurse on the ortho-

paedic wards.

3.3 | Patient characteristics

The mean age of the 50 patients was 66 years (SD 10.3); 26 (52%)

had TKR surgery and 24 (48%) patients had THR surgery. The char-

acteristics of patients on both wards were similar, except that

patients on Ward B were significantly younger (mean 62 (SD 12.0)

years than Ward A (mean 68 years (SD 7.6), p = .035) (Table 2). On

admission, 34 (68%) of patients were assessed as having a high risk

of developing VTE during their acute care admission, however, the

admission VTE risk assessment was not documented in 14 (28%)

patients. Patients were assessed for implementation of mechanical

and pharmacological interventions DVT prophylaxis on days one and

three. All patients were receiving both mechanical and pharmacologi-

cal DVT prophylaxis, 85% of patients were wearing thromboembolic

deterrent stockings, and all patients were encouraged to perform

regular range of motion exercises. Thirty-five (70%) of patients were

on the low molecular weight heparin, enoxaparin sodium (ClexaneTM),

and 15 (25%) were on dalteparin (FragminTM). In addition, eight (16%)

patients were also on regular daily aspirin therapy.

3.4 | Inter-rater agreement

Clinical assessment for the presence of symptoms indicative of pos-

sible DVTs was performed on Days 1, 3 and 5 postoperatively, the

number of patients included decreased over time, largely due to

early discharge from acute care. At total of 114 paired assessments

O’BRIEN ET AL. | 1807

were performed: On Day 1, n = 50; Day 2, n = 46; and Day 3,

n = 18.

The level of agreement between the nurse expert and the surgi-

cal nurses for individual signs and symptoms assessed using the

“STOPDVTS tool” varied between 44%–94% (Table 3). Overall agree-

ment was highest for data collected for the assessment of the pres-

ence of “shortness of breath” (74%) and “tachycardia” (83%) and the

presence of swollen veins or varicosities (77%) and discoloration of

the skin on the affected limb (77%). In contrast, percentage agree-

ment for subjective signs such as the presence of “increased limb

swelling” (55%) and “altered limb skin temperature” (63%) was lower.

The level of agreement between the surgical nurses and nurse

expert assessments increased slightly from Day one to Day three for

symptoms such as “swelling” (increased from 46%–70% agreement)

and the presence of areas of altered limb skin temperature (in-

creased from 68%–76% agreement). The lowest level of agreement

was found for the presence of “increased pain” in the limb (agree-

ment ranged from 44%–56%) and assessing whether the patient was

“more immobile than expected” (agreement ranged from 44%–63%).

Compared to the nurse expert, surgical ward nurses appeared to

under-report signs such as increased pain, increased swelling and

limb skin temperature changes (Figure 1). In 33% of assessments,

the expert nurses assessed the patient as having “increased pain”

and the surgical nurse did not, and in 9%, the surgical nurse identi-

fied increased pain and the expert nurses did not. In 29% of assess-

ments, the expert nurse identified increased swelling of the affected

limb and the surgical nurses did not and in 10%, the surgical nurses

identified swelling and the expert nurses did not. Limb skin tempera-

ture changes were identified as present by the expert nurses in 22%

of assessments in which the surgical nurses did not identify

increased pain.

3.5 | Documentation of the STOPDVTS tool in care records

The ward nurses documented findings of their assessments in the

patient’s medical record at least once per shift; however, specific

findings associated with signs or symptoms of a possible DVT were

not well documented (Figure 2). On Ward A, only three (19%) of the

16 patients who had swelling identified on assessment and none of

the six patients who had evidence of increased warmth or heat in

their limb/s had it documented in the nursing notes within the medi-

cal record. Similarly, on ward B, only one (13%) of eight patients

with evidence of warmth in their limb/s and only one (5%) of 18

patients who had swelling present on Days 1–3 had this docu-

mented in their health record.

TABLE 2 Patient characteristics on admission

Ward A N = 27

Ward B N = 23

Total N = 50

Age (years), Mean (SD) 67.9 (7.6) 61.6 (12.0) 66.2 (10.3)

Sex (%)

Male 17 (63) 8 (35) 25 (50)

Female 10 (37) 15 (65) 25 (50)

Type of surgery

Hip, n (%) 10 (37) 16 (70) 26 (52)

Knee, n (%) 17 (63) 7 (30) 24 (48)

Risk factors for the development of venous thromboembolism (%)

Age >65 years 17 (63) 10 (43) 27 (54)

Lower limb immobility 7 (26) 4 (17) 11 (22)

Obesity 1 (4) 0 1 (2)

Hyperlipidaemia 2 (7) 0 2 (4)

Diabetes 4 (15) 5 (22) 9 (18)

Chronic heart failure 1 (4) 0 1 (2)

Past history Deep Vein

Thrombosis or PE

2 (7) 1 (4) 3 (6)

TIA/stroke day 1 1 (4) 0 1 (2)

Past history smoking 1 (4) 5 (22) 6 (12)

Hypertension 11 (41) 15 (65) 26 (52)

Atrial fibrillation 1 (4) 2 (17) 3 (6)

Oral contraception 0 1 (4) 1 (2)

Varicose veins 5 (19) 0 5 (10)

Active cancer 2 (7) 0 2 (4)

History cancer 2 (7) 2 (17) 4 (8)

Venous thromboembolism (VTE) risk assessment form (%)

Completed 17 (63) 8 (35) 25 (50)

Incomplete 4 (15) 12 (52) 16 (32)

No form 6 (22) 3 (13) 9 (18)

VTE risk stratification (%)

High 20 (74) 14 (61) 34 (68)

Low 0 2 (9) 2 (4)

Not done 7 (26) 7 (30) 14 (28)

TABLE 3 Percentage agreement between the expert nurse and surgical nurses

Day 1 (%) Day 3 (%) Day 5 (%)

Average agreementa

Swelling 23 (46) 31 (70) 9 (50) 55

Shortness of

breath

41 (82) 39 (78) 11 (62) 74

Touch

(hot/cold)

34 (68) 35 (76) 8 (44) 63

Tachycardia 44 (88) 41 (89) 13 (72) 83

Pain 28 (56) 24 (52) 8 (44) 51

Discoloration 36 (72) 34 (74) 12 (67) 71

Veins (presence of

varicosities)

38 (76) 40 (87) 12 (67) 77

Still/immobile in

bed

26 (52) 29 (63) 8 (44) 53

No of patients

assessed per day

n = 50 n = 46 n = 18 N = 114

aAverage agreement over the three days that the assessments were per-

formed.

1808 | O’BRIEN ET AL.

3.6 | Patient outcomes

The outcomes of all 50 patients were followed up at 21 days post-

surgery using the hospital database to identify if there were any re-

presentations associated with possible DVTs. Seven (14%) patients

(TKR n = 4 and THR n = 3) presented to the emergency department

between 4 and 9 days postoperatively and were investigated for DVT

using ultrasound diagnostics. Three of these patients had been identi-

fied as high risk using the hospital VTE screening tool, the other four

had either an incomplete VTE risk assessment documented or it was

absent from the care record. All seven patients had symptoms consis-

tent with possible DVT identified using the STOPDVTS tool (when

completed by the expert nurse): seven (100%) had increased pain,

three (43%) were less mobile than expected and one (14%) had

changes in skin temperature in their calf. Only one of these patients

was diagnosed with a DVT on ultrasound.

4 | DISCUSSION

This study has demonstrated gaps in the translation of evidence into

practice for VTE risk screening and clinical assessment of possible

DVT in high-risk postoperative arthroplasty patients. Despite an

implementation strategy well supported by a structured clinical edu-

cation programme supported by the nursing leadership, there were

gaps in documentation of the admission VTE assessment and docu-

mentation of changing clinical signs indicating possible DVT in the

medical record. Despite structured education in how to assess post-

operative patients using the STOPDVTs tool, this pilot study found

only moderate levels of agreement between expert and surgical

nurse assessors on the presence or absence of signs indicating

possible DVT. The low levels of agreement between the expert and

surgical nurses when using the STOPDVTS tool suggested surgical

nurses were underestimating the presence of clinical findings.

Overall, the total STOPDVTS clinical scores were higher when

the assessment was performed by a nurse expert compared with the

surgical nurses. The low to moderate levels of agreement on some

components of the STOPDVTS assessment tool suggest that the surgi-

cal nurses consistently underestimated the presence of several symp-

toms associated with possible DVT. The highest levels of agreement

were found for items captured in the patient past history such as the

presence of varicose veins and items related to objective measures

such as vital signs. The lowest levels of agreement were found for

items related to physical examination for local symptoms in the

affected limb (e.g., presence of swelling, changes in skin temperature,

skin discoloration) and worsening limb pain. The low levels of agree-

ment for these local signs and symptoms confirm previous work that

has found that clinical assessment for possible DVT (particularly in the

early stages) is highly subjective and even well-trained clinicians may

miss early clinical signs (Falck-Ytter et al., 2012; Songwathana et al.,

2011; Strijkers et al., 2011; Tan et al., 2009; Tenna et al., 2012).

It appears that the surgical nurses did not link together combina-

tions of signs and symptoms as “potentially clinically significant” and

possibly associated with DVT. Despite nurses detecting signs such

as increased limb swelling in their patients, this was documented in

the patient medical record only 19% of the time. These findings are

cause for concern because if surgical nurses overlook clinically signif-

icant combinations of symptoms suggestive of early DVT, particularly

in a high-risk patient group, they may fail to escalate care for

patients that could benefit from further investigation and treatment

(Ali & Young, 2012; Bacon, 2013; Goodacre et al., 2005; Song-

wathana et al., 2011; Wells & Ginsberg, 1995). This problem is

FIGURE 2 Percentage of patients with symptoms identified and the percentage recorded in the medical record on Day 1 postoperatively [Colour figure can be viewed at wileyonlinelibrary.com]

O’BRIEN ET AL. | 1809

compounded by the well-known problem (Gatson et al., 2012;

Roberts et al., 2013; Yin & Shan, 2015) of poor documentation of

VTE risk assessments.

There are a number of plausible explanations for these findings.

First, postoperative orthopaedic patients, in particular those with hip

and knee replacement surgery, are expected to have some signs and

symptoms similar to those found with the presence of DVT, in par-

ticular swelling of the lower limb and increased limb pain (Maletis

et al., 2012; Meetoo, 2010). It is possible the nurse expert was

assessing whether symptoms such as “swelling” were present; in

contrast, the surgical nurses may have implicitly interpreted their

findings in terms of whether this sign was worse than they expected

(or previously assessed) for the individual patient (Kaur et al., 2012).

To address this issue, the current dichotomous response variable

(present/absent) for each item on the STOPDVTS tool was changed to a

multiple response nominal scale that asks the surgical nurse to evaluate

whether each sign or symptom was absent, mild (or as expected), mod-

erate or severe. Such an approach could assist nurses differentiate “ex-

pected” findings in patients who have had surgery from those

“unexpected” more significant findings that warrant further investiga-

tion (Bauersachs, 2012). This revision is expected to address a key limi-

tation impacting the usability of the STOPDVTS tool and may help

nurses to think critically about the clinical significance of the combina-

tions of changes they observe in their patients at initial and subsequent

assessments. Further evaluation of a revised tool will be required to

evaluate whether such a revision also improves the face and content

validity of the tool.

Only one patient enrolled in this study developed a DVT; there-

fore, we were not able to evaluate if high scores calculated using the

STOPDVTS tool were predictive of the presence of a DVT (Hotoleanu

et al., 2010; Penaloza et al., 2006). It was, however, observed that

patients who presented to the emergency department within three

weeks of discharge with symptoms of possible DVT were also symp-

tomatic on the STOPDVTS scores obtained during their admission

(Maletis et al., 2012). It is therefore possible that patients with high

scores on the STOPDVTS tool should be referred for further investiga-

tions prior to hospital discharge. A larger study would need to be con-

ducted to assess the predictive value of the STOPDVTS assessment

score in early detection of DVT in postoperative patients.

Strengths of this study included a comprehensive review of cur-

rent literature and expert consensus supported content validity of

the STOPDVTS clinical assessment tool and; an evaluation of its

introduction onto the wards highlighted the importance of DVT

assessment to the nursing staff and provided them with the oppor-

tunity to undertake training that was not otherwise provided.

As this was a small single site pilot study of only 50 patients, the

ability to generalise findings to other health services or patient popu-

lations is limited. Further, use of a nonrandomised sample and obser-

vational methods means that we are unable to compare our results

to a control group of staff who did not receive the educational inter-

vention. This attempt to provide a valid and reliable standard clinical

assessment tool was hampered by the poor level of agreement

between the expert and the surgical nurses suggesting either limited

face validity or limited skill of the participating nurses (Kaur et al.,

2012; Meetoo, 2010), hence there is a need for further revision of

STOPDVTS the clinical assessment tool.

4.1 | Summary and future directions

This study identified a possible risk to patient safety related to under-

recognition of the signs and symptoms of possible DVT by surgical

nurses in postoperative patients. The findings demonstrate the feasi-

bility of using a protocol for consistent screening for possible DVT in

the postoperative period and provide the foundation for development

of and future research to (1) revise the STOPDVTS clinical assessment

tool, (2) examine tool reliability and validity with nurses caring for

high-risk patient populations. These initiatives will be embedded into

process improvement activities to ensure that these changes to care

processes are sustained over the longer-term.

ACKNOWLEDGEMENT

The authors of this manuscript would like to acknowledge the Mr

Laurie Simpson (Thoracic Surgeon) for his expertise, training and

mentoring of the clinical team and passionate interest in this topic.

CONTRIBUTIONS

Study design: AOB, BR, MB, AH; data collection and analysis: AOB,

BR, BW, MB, AH; and manuscript preparation: AOB, BR, MB, AH.

ORCID

Bernice Redley http://orcid.org/0000-0002-2376-3989

Anastasia F Hutchinson http://orcid.org/0000-0002-0014-689X

REFERENCES

Ali, M., & Young, V. (2012). Shifting DVT identification into primary care.

British Journal of Healthcare Management, 18, 211–214. https://doi.

org/10.12968/bjhc.2012.18.4.211

Australian Commission on Safety and Quality in Health Care (ACSQHC).

(2011). National Safety and Quality Health Service Standards Using

evidence to reduce blood clots – VTE prevention is a priority. Sydney.

Retrieved from http://www.safetyandquality.gov.au/media_releases/

using-evidence-to-reduce-blood-clots-vte-prevention-is-a-priority/

Autar, R. (2003). The management of deep vein thrombosis: The Autar

DVT risk assessment scale re-visited. Journal of Orthopaedic Nursing,

7, 114–124. https://doi.org/10.1016/S1361-3111(03)00051-7

Bacon, S. (2013). Looking again at VTE 1: Assessment and diagnosis.

[Peer Review]. Practice Nursing, 24, 454–458. https://doi.org/10.

12968/pnur.2013.24.9.454

Baser, O., Supina, D., Sengupta, N., Wang, L., & Kwong, L. (2010). Impact

of postoperative venous thromboembolism on medicare recipients

undergoing total hip replacement or total knee replacement surgery.

American Journal of Health-System Pharmacy, 67, 1438–1445.

https://doi.org/10.2146/ajhp090572

Bauersachs, R. (2012). Clinical presentation of deep vein thrombosis and

pulmonary embolism. Best Practice and Research Clinical Haematology,

25, 243–251. https://doi.org/10.1016/j.beha.2012.07.004

1810 | O’BRIEN ET AL.

Bottaro, F., Ceresetto, J., Emery, J., Bruetman, J., Emery, N., Pellegrini, D.,

& Bullorsky, E. (2012). Cross-sectional study of adherence to venous

thromboembolism prophylaxis guidelines in hospitalized patients. The

Thrombo-Brit study. Thrombosis Journal, 10(7), 1–8.

Dahl, O., Gudmundsen, T., & Haukeland, L. (2000). Late occurring clinical

deep vein thrombosis in joint-operated patients. Acta Orthopaedica

Scandinavica, 71(1), 47–50. https://doi.org/10.1080/

00016470052943883

Dewar, C., & Corretge, M. (2007). Interrater reliability of the Wells score

as part of the assessment of DVT in the emergency department:

Agreement between consultant and nurse practitioner. Emergency

Medicine Journal, 25, 407–410.

Emadi, A., & Streiff, M. (2011). Diagnosis and management of venous

thromboembolism: An update a decade into the new millennium.

Archives of Irananian Medicine, 14(5), 341–351.

Falck-Ytter, Y., Francis, C. W., Johanson, N. A., Curley, C., Dahl, O. E.,

Schulman, S., & Colwell, C. W. (2012). Prevention of VTE in orthope-

dic surgery patients: Antithrombotic Therapy and Prevention of

Thrombosis, 9th ed: American College of Chest Physicians Evidence-

Based Clinical Practice Guidelines. Chest, 141(2), e278S–e325S.

https://doi.org/10.1378/chest.11-2404

Gatson, S., White, S., & Misan, G. (2012). Venous Thromboembolism

(VTE) Risk Assessment and Prophylaxis: A comprehensive Systematic

Review of the facilitators and barriers to healthcare worker compli-

ance with clinical practice guidelines in the acute care setting. Joanna

Briggs Institute (JBI) Library of Systematic Reviews, 10, 3812–3893.

Goodacre, S., Sutton, A. J., & Sampson, F. C. (2005). Meta-analysis: The

value of clinical assessment in the diagnosis of deep venous throm-

bosis. Annals of Internal Medicine, 143, 129–139. https://doi.org/10.

7326/0003-4819-143-2-200507190-00012

Grimshaw, J., Eccles, M., Lavis, J., Hill, S., & Squires, J. (2012). Knowledge

translation of research findings. Implementation Science, 7, 1–17.

Hotoleanu, C., Fodor, D., & Suciu, O. (2010). Correlations between clini-

cal probability and Doppler Ultrasound results in the assessment of

deep venous thrombosis. Medical Ultrasonography, 12, 17–21.

Kanchanabat, B., Stapanavatr, W., Meknavin, S., Soorapanth, C.,

Sumanasrethakul, C., & Kanchanasuttirak, P. (2011). Systematic

review and meta-analysis on the rate of postoperative venous throm-

boembolism in orthopaedic surgery in Asian patients without throm-

boprophylaxis. British Journal of Surgery, 98, 1356–1364. https://doi.

org/10.1002/bjs.7589

Kaur, M., Yadav, K., Yadav, V., Gupta, B., Kamble, S., & Misra, M. (2012). Deep

Vein thrombosis (DVT) Prophylaxis: Awareness or ignorance amongst

staff personnel. International Journal of Nursing Education, 4, 32–34.

Kitson, A., Rycroft-Malone, J., Harvey, G., McCormack, B., Seers, K., &

Titchen, A. (2008). Evaluating the successful implementation of evi-

dence into practice using the PARiHS framework: Theoretical and

practical challenges. Implementation Science, 3, 1–12. https://doi.org/

10.1186/1748-5908-3-1

Maletis, G., Inacio, M., Reynolds, S., & Funahashi, T. (2012). Incidence of

symptomatic Venous Thromboembolism after elective knee arthro-

scopy. The Journal of Bone and Joint Surgery America, 95, 714–720.

https://doi.org/10.2106/JBJS.J.01759

Maynard, G. A., Morris, T. A., Jenkins, I. H., Stone, S., Lee, J., Renvall, M.,

. . . Schoenhaus, R. (2010). Optimizing prevention of hospital-acquired

venous thromboembolism (VTE): Prospective validation of a VTE risk

assessment model. Journal of Hospital Medicine, 5, 10–18. https://doi.

org/10.1002/jhm.562

Meetoo, D. (2010). In too deep: Understanding, detecting and managing

DVT. British Journal of Nursing, 19, 1021. https://doi.org/10.12968/

bjon.2010.19.16.78188

National Health and Medical Research Council (NHMRC). (2011). Stop

the Clot: Integrating VTE prevention guideline recommendations into

routine hospital care (3rd ed.). Melbourne Research Council. Available

from www.nhmrc.gov.au. NHMRC reference code: CP134.

National Health and Medical Research Council (NHMRC) (2009). Clinical

Practice Guideline for the prevention of venous Thromboembolism (deep

vein thrombosis and pulmonary embolism) in patients admitted to Aus-

tralian Hospitals. Melbourne, Vic.: Council NHaMR.

Penaloza, A., Laureys, M., Wautrecht, J., Lheureux, P., & Motte, S. (2006).

Accuracy and safety of pretest probability assessment of deep vein

thrombosis by physicians in training using the explicit Wells clinical

model. Journal of Thrombosis and Haemostasis, 4, 278–281. https://d

oi.org/10.1111/j.1538-7836.2005.01740.x

Righini, M., & Bounameaux, H. (2007). Venous thrombosis: Risk factors

and management. Herz, 32, 27–34. https://doi.org/10.1007/s00059-

007-2931-1

Roberts, L., Porter, G., Barker, R., Yorke, R., Bonner, L., Patel, R., & Arya,

R. (2013). Comprehensive venous thromboembolism prevention pro-

gramme incorporating mandatory risk assessment reduces the inci-

dence of hospital-associated thrombosis. Chest, 144, 1276–1281.

https://doi.org/10.1378/chest.13-0267

Schiff, R., Kahn, S., Shrier, I., Strulovitch, C., Hammouda, W., BPharm, E.,

& Zukor, D. (2005). Identifying orthopedic patients at high risk for

venous thromboembolism Despite Thromboprophylaxis. Chest, 128,

3364–3371. https://doi.org/10.1378/chest.128.5.3364

Songwathana, P., Promlek., K., & Naka, K. (2011). Evaluation of a clinical

nursing practice guideline for preventing deep vein thrombosis in crit-

ically ill trauma patients. Australasian Emergency Nursing Journal, 14,

232–239. https://doi.org/10.1016/j.aenj.2011.09.002

Strijkers, R., Cate-Hoek, A., Bukkems, S., & Wittens, C. (2011). Manage-

ment of deep vein thrombosis and prevention of post-thrombotic

syndrome. British Medical Journal, 343, 949–953.

Tabei, L. E., Holtz, E., Sch€urer-Maly, C., & Abholz, H. (2012). Accuracy in

diagnosing deep and pelvic vein thrombosis in primary care.

Deutsches €Arzteblatt International, 109, 761–766.

Tan, M., Van Rooden, C. J., Westerbeek, R. E., & Huisman, M. V. (2009).

Diagnostic management of clinically suspected acute deep vein

thrombosis. British Journal of Haematology, 149, 347–360. https://doi.

org/10.1111/j.1365-2141.2009.07732.x

Tenna, A., Kappadath, S., & Stansby, G. (2012). Diagnostic tests and

strategies in venous thromboembolism. Phlebology, 27, 43–52.

https://doi.org/10.1258/phleb.2012.012s35

Wells, P., & Anderson, D. (2013). The diagnosis and treatment of venous

thromboembolism. Hematology, 2013, 457–463. https://doi.org/10.

1182/asheducation-2013.1.457

Wells, P., Borah, B., Sengupta, N., Supina, D., MsDonald, H., & Kwong, L.

(2010). Analysis of venous thromboprophylaxis duration and out-

comes in orthopedic patients. The American Journal of Managed Care,

16, 857–866.

Wells, P. S., & Ginsberg, J. S. (1995). DVT and pulmonary embolism:

Choosing the right diagnostic tests for patients at risk. Geriatrics, 50

(2), 29–32, 35–36.

Wells, P. S., & Scarvelis, D. (2006). Diagnosis and treatment of deep-vein

thrombosis. Canadian Medical Association Journal, 175, 1087–1092.

Yin, H.-Z., & Shan, C.-M. (2015). The effect of nursing intervention based

on Autar scale results to reduce deep venous thrombosis incidence in

orthopaedic surgery patients. International Journal of Nursing Sciences,

2, 178–183. https://doi.org/10.1016/j.ijnss.2015.04.003

How to cite this article: O’Brien A, Redley B, Wood B, Botti

M, Hutchinson AF. STOPDVTs: Development and testing of a

clinical assessment tool to guide nursing assessment of

postoperative patients for Deep Vein Thrombosis. J Clin Nurs.

2018;27:1803–1811. https://doi.org/10.1111/jocn.14329

O’BRIEN ET AL. | 1811