Table 1 Included studies

Trompeter et al., 2010

Retrospective Diagnostic validity

Orthopaedic clinic (United Kingdom)

Triage of patients for orthopaedic consultation

Knee soft tissue or sports injuries

100

1- Comparison of diagnostic accuracy to arthroscopy for:

1- Sensitivity

a. 68.1%

b. 90.7%

a. APP

Specificity

b. Orthopeadic surgeon

a. 66.6%

b. 71.4%

Comparison in number of incorrect diagnosis:

17/50 for APP compared to 9/50 for surgeon (p < 0.07)

2- Identification of surgical candidates

2- Number of correctly selected surgical candidates

a. APP

a. 47/50

b. Orthopeadic surgeon

b. 43/50

No significant differences between providers (p = 0.20)

MacKay, et al., 2009

Inter-rater agreement

Orthopaedic clinic (Canada)

Triage of patients for orthopaedic consultation and treatment recommendations (rehabilitation interventions)

Hip and knee arthritis

62

Agreement between APPs and Orthopaedic surgeons:

1- Appropriateness to be seen by surgeon

1- Level of agreement κ = 0.69

Observed agreement 91.8%

2- Identification of TJA surgical candidates

2- Level of agreement κ = 0.70

Observed agreement 85.5%

Aiken et al., 2008

Inter-rater agreement

Orthopaedic clinic (Canada)

Triage of surgical candidates for TJA and treatment recommendations (rehabilitation, medication, ordering tests, referral to other providers)

Hip and knee arthritis**

38

Agreement between an APP and an Orthopaedic surgeon:

1- Identification of TJA surgical candidates

1- Observed agreement 100%

2- Surgical urgency using the WCWL-HKPT tool

2- Observed agreement 64%

3- Treatment recommendations

3- Level of agreement κ = 0.68

Aiken and McColl, 2008

Diagnostic validity/Inter-rater agreement

Orthopaedic clinic (Canada)

Diagnosis and treatment recommendations (rehabilitation, medication, ordering tests, referral to other providers, and to surgery)

Shoulder or knee musculoskeletal impairments

24

Agreement between an APP and an Orthopaedic surgeon:

1- Diagnostic agreement

1- Level of agreement for knee impairments κ = 0.69

Observed agreement for knee and shoulder impairments 90%

2- Treatment recommendations

2- Level of agreement κ = 0.52–0.87

Observed agreement 90%

Diagnostic accuracy of APP compared to MRI:

3- Diagnostic agreement

3- APP accuracy to MRI 75%

Orthopedic surgeon accuracy to MRI 75%

O’Donoghue and Hurley-Osing, 2007

Diagnostic validity

Physiotherapy hospital department (Ireland)

Diagnosis of new patients referred by the emergency department

Acute knee injury, of less than three weeks duration

42

Diagnostic accuracy of an APP compared to MRI

All knee derangements, PPV = 73,2

ACL tear, PPV = 90,4

Meniscal tear PPV = 55.5

Moore, J. H., 2005

Retrospective Diagnostic validity

Military hospital clinic (United States)

Primary care practitioner (rehabilitation, medication, ordering tests, referral to other providers and to surgery)

Musculoskeletal complaints of the spine or extremities

560

Comparison of diagnostic accuracy to MRI for:

Observed diagnostic accuracy:

a. 74.5% (108/145)

b. 80.8% (139/172)

a. APPs

c. 35.4% (86/243)

b. Orthopeadic surgeons

Difference in diagnostic accuracy between groups:

c. Other healthcare providers^†

a better than c (P = 0.001)

b better than c (P = 0.001)

No differences between a and b (P > 0.05)

Dickens, et al., 2003

Diagnostic validity/inter-rater agreement

Orthopaedic clinic (United Kingdom)

Diagnosis and triage of surgical candidates for arthroscopy

Knee impairments excluding severe osteoarthritis

50

Agreement between APPs and an Orthopaedic surgeon:

1- Observed agreement 76.5%

1- Diagnostic agreement

2- Diagnostic accuracy to arthroscopy:

Sensitivity (range depending on pathology)

2- Diagnostic accuracy to arthroscopy for:

a. 43–93%

b. 40–100%

a. APPs

Specificity

b. Orthopeadic surgeons

a. 92–98%

b. 98–100%

Sephton et al., 2010

Prospective observational cohort

Outpatient musculoskeletal clinic (United Kingdom)

Triage of patients for orthopaedic, rheumatology or pain clinic consultations (ordering tests, referral to other providers and to surgery)

Various musculoskeletal conditions

217

Treatment outcomes for patients triaged by APP at 3 months and 12 months following care (no control group):

Mean improvement in scores and 95%CI from baseline to 3 and 12 months:

1- Pain VAS(/10)

1- 3 m: −0.72 (−1.15 to −0.29)

12 m: −0.80 (−1.31 to −0.29)

2- EQ-5D questionnaire (/1)

2- 3 m: 0.044 (0.001 to 0.086)

12 m: 0.048 (0.003 to 0.093)

3- SF-36 questionnaire (%)

3- 3 m: −0.9% (−6.3 to 4.4)

12 m: −4.9%(−9.9 to 0.1)

4- Perceived improvement-PIVAS scale (%)

4- 3 m : 33% (28 to 38)

12 m: 46% (40 to 51)

5- Deyo and Diehl Satisfaction Questionnaire (%)

Proportion of patients satisfied with care:

5- 94%

Taylor et al., 2010

Prospective non-randomised controlled trial

Three emergency departments—ED (Australia)

Primary care practitioner (rehabilitation, medication and ordering tests)

Peripheral musculoskeletal injury

315

Comparison between first line APP care and usual medical care followed by physiotherapy care for ED consultation:

Differences and 95%CI between APP care and usual care:

Time reduction with APP care:

1- Length of stay (min)

1- 59.5 (38.4 to 80.6) min.

2- Wait time (min)

2- 25.0 (12.1 to 38.0) min.

3- Treatment time (min)

3- 34.9 (16.2 to 53.6) min.

Relative Risks (APP relative to usual care):

4- Proportion of re-presentation to ED at 1 month follow up

4- RR : 1.02 (0.51 to 2.05)

5- Proportion of diagnostic imaging referrals

5- RR : 0.89 (0.78 to 1.02)

Proportion of patient satisfied and relative risk (APP relative to usual care):

6- APP care : 85%

Usual care: 82%

6- Patient satisfaction

RR: 1.03 (0.94 to 1.15)

Ball and Walton, 2007

Retrospective observational cohort

Emergency department (United Kingdom)

Primary care practitioner (rehabilitation, medication and ordering tests)

Closed musculoskeletal injuries to the upper or lower extremities, including fractures

643

Comparison between APPs, nurse practitioners and physicians (senior house officers, middle grade doctors and consultants):

1- No differences between providers (p = 0.17)

2- No differences between providers (p = 0.99)

1- Proportion of ordered X-rays

3- APP gave more advice (p < 0.007)

2- Proportion of positive X-rays

APP prescribed fewer assistive devices (p < 0.001)

APP referred more patients to physiotherapy (p < 0.001)

3- Soft tissues injury treatment recommendations

Physicians prescribed more medication than other providers (p < 0.001)

McClellan et al., 2006

Prospective quasi- experimental cohort

Emergency department (United Kingdom)

Primary care practitioner (rehabilitation, medication and ordering tests)

Patients with peripheral soft tissue injuries and associated fractures

102^‡

Comparison between APPs, nurse practitioners (NP) and physicians on treatment outcomes for patients with ankle injuries only at 4 or 16 weeks:

784^°

Mean Wait and consultation times comparisons:

1- Mean wait time for consultation (min.)

1- APPs: 43 min., NPs: 55 min., Physicians: 80 min.

APP significantly shorter wait time than NP and physicians (p < 0.05)

2- Mean consultation time (min.)

2- APPs: 25 min., NPs: 15 min., Physicians: 20 min.

No significant differences in consultation time (p > 0.05)

3- Pain VAS (/10)

Outcome of treatment for patients with ankle injuries only at 4 weeks:

4- Function VAS (/10)

3- No significant differences between providers (p > 0.05)

5- SF-36 (%)

4- No significant differences between providers(p > 0.05)

Comparison between APPs, nurse practitioners and physicians care for all patients and type of injuries:

5- No significant differences between providers (p > 0.05)

6- Patient satisfaction (%)

Proportions of patient satisfied with care (patient who strongly agreed to question: Overall I was satisfied with the treatment received):

6- APPs: 54.5% NPs: 38.9%, Physicians: 35.6% (p = 0.048)

Richardson et al. 2005

RCT and cost consequence analysis

Emergency department (United Kingdom)

Primary care practitioner (rehabilitation, medication and ordering tests)

Patients with semi or non-urgent musculoskeletal conditions

766

Comparison between APP care and usual care by emergency physician on treatment outcomes at 6 months:

Difference and 95%CI for days to return to usual activities or work:

1- Return to usual activities (days)

1- 12.5 added days for APP care. APP care marginally longer than usual care (p = 0.07)

2- Return to work (days)

2- 1 added day for APP care (−3.0 to 1.0). No differences between providers (p > 0.05)

Difference in proportions of patient satisfied with care and 95%CI:

3- Satisfaction with care

3- 74% for usual care and 89% for APP care : 15% difference (9 to 21%)

Economic analysis

4- Direct costs to healthcare system

4- No differences in costs between the two types of care (p > 0.05)

5- Direct costs to patients

5- No differences in costs between the two types of care (p > 0.05)

6- Indirect costs (productivity loss)

6- No differences in costs between the two types of care (p > 0.05)

Daker-White et al., 1999

RCT and cost minimisation analysis

Orthopaedic clinic (United Kingdom)

Primary care practitioner (rehabilitation, medication, ordering tests, referral to other providers and to surgery)

Patients with musculoskeletal complaints

481

Comparison between APP care and usual care by orthopeadic surgeons in training (UK junior doctors):

Treatment outcomes at a mean 5.6 months follow-up:

No significant differences between providers for outcomes 1 to 8 (p > 0.05)

Treatment outcomes at a mean 5.6 months follow-up:

1- Pain VAS (/10)

Use of health services:

2- Oswestry Disability Index (%)

3- St-Michael's (48-0)

9- Significant difference in the proportion of patients with no test ordered (p < 0.01): 14.7% for surgeons and 47.5% for APP

4- WOMAC (0–96)

5- Perceived handicap (DRP)

6- SF-36 (%)

7- Psychological status (HADS)

Significant difference in the proportion of patients with X-rays ordered (p < 0.01): 41.4% for surgeons and 13% for APP

8- Self-efficacy

Use of health services

10- Significant difference in the proportion of patients who received advice and reassurance (p < 0.01): 32.5% for surgeons and 58.9% for APP

9- Use of diagnostic tests for consult

10- Treatment recommendations

Satisfaction with care

Significant difference in the proportion of patients who received Intra-muscular injections (p < 0.01): 3.9% for surgeons and 0.5% for APP

11- Patients

12- Referring general practitioners

Economic analysis

Significant difference in the proportion of patients who were referred for surgery (p < 0.01): 17% for surgeons and 7.1% for APP

13- Direct costs to patients

14- Direct costs to healthcare system (NHS)

Satisfaction with care for patients and referring GP

11- Satisfaction scores and 95%CI:

Staff communication/attitudes (scale from 19–95) 4.6 points significant difference (2.2 to 6.8) favoring APP care

Perceived treatment quality (scale from 13–65) 3.0 points significant difference (1.3 to 4.9) favoring APP care.

Facilities (scale from 5–25) 0.9 point significant difference (0.3 to 1.7) favoring APP care.

12- No significant differences between providers (p > 0.05)

Direct costs differences

13- No differences in costs between the two types of care (p > 0.05)

14- Significant difference in direct hospital costs (p < 0.01):

£498.38 for surgeon care and £255.55 for APP care.

Hockin and Bannister, 1994

Retrospective observational cohort

Orthopaedic clinic (United Kingdom)

Primary care practitioner (rehabilitation, orthotic, injection, ordering tests, referral to other providers and to surgery)

Patients with musculoskeletal complaints

189

Patient self reported global perception of improvement (%):

1- At the end of treatments by APP

1- 71% of patients improved by more than 40% on scale of improvement.

2- Comparison of type of APP treatment and proportion of patients who improved:

2- More patients reported improvement with orthotics or injections than with advice and physiotherapy or surgery and referrals to other medical providers. (p < 0.05)

Kennedy et al., 2010

Cross-sectional observational study

Orthopaedic clinic (Canada)

Follow-up care after hip and knee arthroplasty

Hip and knee arthroplasty patients

123

Comparison of patients satisfaction measured by the modified VSQ-9 questionnaire:

Satisfaction score

a. APP led follow-up clinic

a. 89.8%

b. Orthopaedic surgeon led follow-up clinic

b. 87.6%

No significant differences between providers (p = 0.34)

Campos Ayling et al. 2002

Cross-sectional observational study

Paediatric rheumatology clinic (Canada)

Review and manage independently pre-selected patients and refer to rheumatologist when tests and medication are needed

Pediatric patients with Juvenile Idiopathic Arthritis

358

Comparison of patients satisfaction measured by the modified GHAA questionnaire

Summary satisfaction score (5 point scale):

a. APPs led clinic

a. 4.0 ±0.7

b. Rheumatologists led clinic

b. 4.0 ±0.7

No significant differences between care models (P > 0.05)