are we using support workers effectively? the relationship between patient and team characteristics...

Are we using support workers effectively? The relationship between patient and

team characteristics and support worker utilisation in older people’s

community-based rehabilitation services in England

Anna Moran PhD1, Susan Nancarrow BSc (Podiatry) MSc (Research) PhD

2, Pamela Enderby MBE PhD MSc FRCSLT

DSc (Hons)3 and Mike Bradburn MSc

3

1School of Community Health, Charles Sturt University, 2Health and Human Sciences, Southern Cross University,

Australia, and 3School of Health and Related Research, University of Sheffield, UK

Accepted for publication 2 February 2012

CorrespondenceAnna MoranSchool of Community HealthCharles Sturt UniversityPO Box 789, AlburyNSW 2640, AustraliaE-mail: [email protected]

What is known about this topic

d Support workers are utilised toincrease the flexibility and efficiencyof the health and social care work-force.

d Over 80% of community-basedrehabilitation services (CBRS) utilisesupport workers.

d CBRS staff perceive that supportworkers undertake the majority ofdirect patient care.

What this paper adds

d CBRS support workers do notundertake the greatest proportion ofdirect patient care, and the amountof care they provide is associatedwith patient and team factors.

d Less-dependent patients onadmission are more likely to havehigher proportion of supportworker input.

d Targeted utilisation of supportworkers according to patient andteam factors may enable a moreeffective use of the health andsocial care workforce.

AbstractThis research explores the relationship between support worker

utilisation and patient- and team-level characteristics using a prospectivelongitudinal study of 20 older people’s community rehabilitation teams.

Between January and September 2009, 462 Whole Time Equivalent Staff

and 1913 patients participated in the study. Patient dependency, health

status (Therapy Outcome Measures, EQ-5D and Levels of Care tool) and

demographic data were collected alongside detailed staff activity data

for patients during the recruitment period. Multivariate analyses were

used to determine the relationship between the proportion of care deliv-

ered by support workers team and patient variables. Support workersdelivered up to 36% of direct patient care and spent less time per patient

contact (36.0 minutes, SD 37.7, range 0–600) than qualified professionals

(mean time per contact 75.8 minutes, SD 32.9, range 0–334). Less-depen-

dent and female patients had a greater proportion of support worker

input, with a 4.9% increase in face-to-face support worker time for every

0.1 unit increase in EQ-5D (95% CI 0.3–9.6, P = 0.038), while females had

an additional 5.1% of their total contact time with support workers com-

pared with males (95% CI 1.9–8.4, P = 0.002). In an analysis without EQ-5D, older patients had a greater proportion of support worker time

(P = 0.006). Other factors associated with support worker input include

the referral source, intensity of care, usual living arrangements and pro-

portion of support staff in a team. Results indicate that patient- and

team-level factors are a determinant in the proportion of support worker

care delivered to older people in the community, suggesting that it may

be appropriate to introduce a more explicit targeted deployment of sup-

port workers based on patient-level characteristics, which may facilitatea more effective use of qualified practitioner time for more complex,

dependent patients.

Keywords: assistants, community rehabilitation, skill mix, support workers,

workforce

Introduction

Support workers are considered integral to service provi-

sion across many health settings. More than one million

support staff are utilised in public, voluntary and private

sector health and social care organisations across the

United Kingdom (Saks & Allsop 2007). Heralded as

conduits to increasing service capacity and improving

ª 2012 Blackwell Publishing Ltd 1

Health and Social Care in the Community (2012) doi: 10.1111/j.1365-2524.2012.01065.x

workforce efficiency within Community-Based Rehabili-tation Services (CBRS) (Rolfe et al. 1999, Stevenson 2000,

Hart et al. 2005, Stanmore et al. 2005, Stanmore & Water-

man 2007), over 80% of these services now utilise sup-

port workers to deliver care (Nancarrow et al. 2009).

Despite widespread deployment in CBRS, there is little

empirical evidence to demonstrate the cost benefits and

contribution support workers make to the delivery of

care in these settings.For the purpose of this research, the term support

worker refers to all workers who work with profession-

ally qualified staff who may have health or social care

training, such as National Vocational Qualifications, but

who do not have tertiary or equivalent qualifications.

These may include the following: assistants to profes-

sional groups (physiotherapy, occupational therapy,

social work, nursing, etc.), generic assistants (who workacross professional groups), technical instructors, health-

care workers, home helpers ⁄ enablers ⁄ carers and social

services support staff.

There are a number of assumptions about the way

that including support workers in the skill mix may be

used to enhance service outcomes. Managers report that

support workers are utilised for cost-effectiveness, flexi-

bility in working hours and also as a response toresource constraints such as nursing or therapist short-

ages (Thornley 2000). Within CBRS, support workers are

perceived as a means to enable more effective use of

qualified therapists, allowing them to concentrate more

on assessments, case management and complex treat-

ments by allocating prescribed treatment plans to assis-

tants (Russell & Kanny 1998, Steele & Wright 2001,

Mackey 2004, Reid 2004, Taylor & Birch 2004, Stanmore& Waterman 2007). As such, there is a perception

amongst CBRS staff that support workers undertake the

majority of direct patient care with qualified staff under-

taking a more assessment, triage and expert advisory

role (McClimens et al. 2010, Nancarrow et al. 2010).

Many of these assumptions have been translated into

policy directives in the UK. The role of support workers

has been endorsed as a means of increasing the flexibilityand efficiency of the workforce in meeting patient and

service needs (Department of Health 2000a,b, 2001,

2004). Given the current policy directives and assump-

tions, the increasingly ageing population and workforce

and the financial burdens of delivering care, the growth

in utilisation of support workers is likely to continue.

Yet the research findings on support worker skill mix

are varied, indicating that the use of less qualified staffmay not be effective in all situations (Buchan & Dal Poz

2002). In particular, there is a distinct lack of evidence

confirming the assumption that support workers facili-

tate improved workforce efficiency. Jenkins-Clarke &

Carr-Hill (2003), in their analysis of the activity and

workload of 5208 nurses and support staff from 19 hospi-tals in the UK, found that qualified nurses do not spend

more time on direct care when there are more support

staff present or when support staff undertake more time

on overheads ⁄ non-direct care. Furthermore, the authors

found that the speciality of the ward (paediatrics, ortho-

paedics, etc.) and patient severity did not consistently

have a large impact on the division of labour between

support and qualified staff. Overall, the study found thatthere was little difference in the types of tasks under-

taken by any level ⁄ type of staff. In CBRS, Knight et al.(2004) found similar levels of variation in activity among

generic support workers, with some spending nearly a

fifth of their time on administrative duties, while others

spent 95% of their time on therapeutic interventions.

This discrepancy between the perceived and actual

benefits of utilising support workers in CBRS is com-pounded by a lack of evidence to refute or confirm these

perceptions. Services and policy-makers therefore have a

limited evidence base from which to inform decisions

about staffing and delivering care. Given the ageing pop-

ulation and workforce, and increasing utilisation of ser-

vices like CBRS, there is a need to ensure the optimum

efficiency and appropriateness of the workforce in deliv-

ering their services to older people.A recent national survey of 171 CBRS found that 80%

of teams employ at least one whole time equivalent

(WTE) support worker, and on average, employ 10.6

WTE support workers (Nancarrow et al. 2009). However,

there are large variations across teams, and the way that

staff work and the client groups with which they work

are unclear.

The following research therefore aims to identify theproportion of direct care support workers deliver in

CBRS relative to qualified practitioners; to examine

whether or not patient and ⁄or team characteristics are

related to the proportion of care delivered by support

workers; and to examine whether the proportion of care

delivered by support worker is associated with patient

health status.

This research forms part of a larger 3-year mixedmethods research project investigating the impact of

workforce flexibility on the costs and outcomes of older

people’s services (Nancarrow et al. 2010).

Methods

The research involved a prospective longitudinal studyof 20 older people’s CBRS across England. Study partici-

pants included the teams, all staff involved in delivering

services within the teams and a cohort of newly admitted

patients to each team recruited consecutively over a 3-

month period. Data collection took place between Janu-

ary and September 2007.

A. Moran et al.

2 ª 2012 Blackwell Publishing Ltd

Recruitment

In the absence of a national database of older people’s

CBRS, several sources were drawn upon to recruit teams;

The community therapists’ network (n = 173 teams), Pri-

mary Care Trust (PCT) chief executives (n = 484) and sixlocal networks were contacted by post inviting them to

participate in a survey about intermediate care service

organisation and structure (Nancarrow et al. 2009).

Teams were eligible for inclusion if their primary cli-

ent base was people over the age of 65 and where their

clients received a package of care aimed at making

them more independent as classified by descriptive

information derived from a review of CBRS (Nancarrowet al. 2009b). All patients admitted to the team during

the recruitment period were eligible for inclusion. There

were no specific patient exclusion critieria for this

study.

Team staff members were responsible for identifying

and recruiting patients to the study. All staff involved in

the collection of patient-level data attended a half-day

training session in data collection and ethical recruit-ment. National Health Service ethical approval was

obtained in 2006 (06 ⁄ Q1606 ⁄ 132) and from Social Ser-

vices in January 2007.

Sampling

The teams that responded to the larger survey about

intermediate care service organisation and structure

(detailed above) (n = 186) were invited to participate in a

prospective study. Target recruitment was 20 teams,

with a view to represent diversity in terms of rurality,

host organisation (e.g. health- or social services-based)

and staffing models. All teams that expressed an interest

to participate in the prospective study and who were eli-gible for inclusion were recruited by consent of team

managers or leaders. Staff from participating teams were

formally invited to participate in the patient data collec-

tion process and written consent was gained.

A statistically generated power estimate was not cal-

culated for this study as this type of research and the

research questions being asked are exploratory and

novel in design. Team sample size was based on theresults from a scoping study of CBRS (Nancarrow et al.2009) and from a large-scale national evaluation of inter-

mediate care (Barton et al. 2005) where over 1000 patient

records from 10 teams over a 3-month recruitment per-

iod were generated and sound statistical analyses of

patient change in health status was possible from these

records. It was estimated therefore that the inclusion of

twenty teams would allow for around 2000 patientrecords to be collected and that this amount of data

would be sufficient to draw reasonably robust conclu-

sions.

Data collection

Data came from three sources: the Client Record Pack

(CRP), Service Proforma and Workforce Dynamics Ques-

tionnaire (WDQ; Box 1).

Client Record PackFor every patient recruited into the study, all staff mem-

bers completed a CRP, which captured detailed staffing

activity for every contact with the patient, as well as

detailed patient demographic and health dependency

information. This included type of contact provided

(face-to-face or administrative), job title and length of

time spent undertaking duties for or with the client. TheCRP also recorded patient demographics (year of birth,

gender), admission level of care need [using the Levels

of Care (LOC) tool; Box 2], admission and discharge

patient health status and other variables such as the

referring professional ⁄ organisation, usual living arrange-

ments and where care was delivered.

Patients were given an information sheet describing

the study and the way their data would be collected andused, and were given the option to ‘opt out’ of data col-

lection in line with Patient Information Advisory Group

(2006) guidelines.

Patient dependency and outcomes data were mea-

sured using the LOC tool, Therapy Outcome Measure

(TOMs) and the EQ-5D. The LOC tool is a matrix

describing eight possible categories of patient need. It

was used in this study as a proxy for the severity ofpatient illness, and to help identify potential groups of

patients based on their level of service requirement

(Enderby & Stevenson 2000). The TOMs scale is a thera-

pist-rated rehabilitation outcome measure containing

four dimensions: Impairment (degree of severity of dis-

order); Activity (degree of limitation); Social participa-

tion; and Wellbeing (effect on emotion ⁄ level of distress),

with each dimension scored from 0 to 5, including half-points. Lower scores indicate higher levels of severity

(Enderby et al. 2006). The EQ-5D is a patient-completed,

generic quality of life measure used to calculate quality-

adjusted life years (The Euroqual Group 1990). It uses a

single question to assess each of five health domains;

mobility, self-care, usual activities, pain ⁄discomfort and

anxiety ⁄ depression. The EQ-5D has a complex scoring

system, which ranges from 1, which indicates full health,to )0.59 (Dolan 1997).

Service ProformaAt team level, data depicting the WTE staff mix of each

team, were captured using a ‘Service Proforma’ (Nancar-

row et al. 2009), which was completed once at the begin-

ning of the study by the team leader. Of interest was the

proportion of WTE support staff in a team.

Rehabilitation support workers and outcomes


Workforce Dynamics QuestionnaireAt staff level, a WDQ was completed by all staff mem-

bers at the beginning of the prospective study at a train-

ing day held for all participating teams. The WDQ is a

validated, 58-item, Likert scale questionnaire, which is

self-completed by staff members. It explores 11 domains

of workforce flexibility in CBRS. It was developed and

validated in the context of older people’s CBRS to mea-

sure staff outcomes such as team integration and roleflexibility (Nancarrow et al. 2010). It is validated to do so

across teams, within teams and across different disciplin-

ary groups. Scores out of 100 are calculated for each sub-

scale, with 0 representing the worst possible outcome

and 100 representing the best possible outcome on the

scale. Of interest was the relationship between levels ofteam integration and support worker input.

Analysis

Of interest was the relationship between the proportion

of direct care delivered by support staff and a number ofteam- and patient-level variables (Box 1). As such, two

outcomes of interest were derived from staffing activity

data (CRP):

(1) Percentage of face-to-face patient time delivered bysupport workers

(2) Percentage of patient contacts delivered by supportworkers

For each patient, the percentage of time delivered by

support workers relative to qualified staff was defined

as:

Percentage

of time¼

100� Total face-to-face time delivered

by support workers

Total face-to-face time delivered either

by support or qualified workers

A similar algorithm was used to calculate the per-

centage of support worker contacts relative to qualified

staff:

Box 2 The levels of care matrix

Level 0: Patient does not need any intervention

Level 1: Patient needs prevention ⁄ maintenance programme

Level 2: Patient needs convalescence ⁄ respite

Level 3: Patient needs slow stream rehabilitation

Level 4: Patient needs regular rehabilitation programme

Level 5: Patient needs intensive rehabilitation

Level 6: Patient needs specific treatment for individual acute

disabling condition

Level 7: Patient needs medical care and rehabilitation

Level 8: Patient needs rehabilitation for complex profound

disabling condition

Box 1 Independent and dependent variables

Independent variables Source of data Dependent ⁄ outcome variables

Age CRP

Gender CRP

TOMs score admission

Impairment; participation; well-being; activity CRP

EQ-5D at admission CRP

Level of dependence at baseline

Level of care (levels 0–8) CRP

Who referred patient CRP

GP; Self ⁄ informal ⁄ carer ⁄ friend ⁄ family;

Community nurse; Social worker; Allied

health professional; A&E ⁄ acute care ward;

Community hospital; Other Intermediate

Care Team ⁄ ward; Other

1. Proportion of face-to-face

patient time delivered by support

workers

Usual living arrangements CRP

Own home, alone; Own home, with other(s);

With relative; Residential ⁄ nursing home;

Sheltered housing; Other

2. Proportion of face-to-face

contacts delivered by support

workers

Where is care being delivered CRP

Own ⁄ relative home; Hospital; Community; Other

Intensity of care CRP

Avg. no. minutes contact per day

Team mean score for integration WDQ

Team mean score for role flexibility WDQ

Size of team Service Proforma

% of support staff in team Service Proforma

CRP, Client Record Pack; TOMs, Therapy Outcome Measures; WDQ, Workforce Dynamics Questionnaire.

A. Moran et al.


Percentage

of contacts¼

100�Number of face-to-face contacts

delivered by support workers

Number of face-to-face contacts delivered

either by support or qualified workers

Factors associated with support worker contribution tocareThe analysis strategy for all covariates was as follows:

univariate analyses were undertaken using a general-

ised least squares random-effects analysis clustering

on team. All covariates (Box 1) were then put into a

multivariate model, and those which had little impact

on any of the outcomes (P > 0.2) were manuallyremoved. The Stata statistical software (StataCorp LP,

College station, TX, USA) version 11.1 was used for

all data manipulation and analyses. Unstandardised

regression coefficients are presented, and indicate a

change in percentages: for example, females, on aver-

age, have an additional 4.5% of their total contact time

with support workers compared with males.

Missing dataMissing data were addressed in the following ways:The TOMs and EQ-5D scores were left blank if miss-

ing, and omitted from the analyses. The reported

results include only data that were complete. We ver-

ified the staff types by cross-checking patient-level

staff activity data against the service proforma. Miss-

ing data with respect to staff activity data (CRP)

were calculated separately for each team using the

average time spent by each type of practitioner ineither ‘face-to-face’ or ‘administration’. Staffing models

and assumptions were verified with team leaders,

where necessary. Where one or more items of data

were missing from any WDQ domain, we imputed

the individual mean score for that domain, unless all

data were missing, in which case, the item was not

calculated (Shrive et al. 2006).

Results

Expressions of interest were received from 27 teams to

participate in the prospective study and 20 were

included in the study. Complete CRPs and service pro-

formas were received for 1882 patients from 20 teams.

Complete staffing data were received for 1708 patients

with 298 WDQs received from staff across 20 teams.

Participating teams

Contextual and staffing data for participating teams

are summarised in Table 1. The most common type

of practitioner employed was support workers (95%),followed by physiotherapists (90%) and occupa-

tional therapists (90%). The mean percentage ofsupport workers per team was 41.1% (SD 14.9, range

0–82.4).

Patient characteristics

Of the 1882 patients for whom we received completepatient record packs, 63% were female with a mean age

of 79.7. Overall mean EQ-5D admission scores were low

across the board, with the average EQ-5D admission

score being 0.4 (SD 0.3, range )0.59 to 1), indicating that

the sample had generally poor health (1 is considered

full health and zero represents a health state considered

by the general population to be equally preferable to

being dead). The mean TOMs admission scores (3.1–3.7,range 0–5 for all domains) demonstrate that patients

admitted to these services also have moderate levels of

disability. Dependency at admission, as measured by the

EQ-5D, and TOMs impairment scores show some

differences between teams (Table 2). One-quarter of

all patients were identified as requiring a regular

rehabilitation programme at admission to the service

(LOC level 4) (Table 2).

Proportion of care delivered by support workers

Across all teams, the proportion of contacts delivered by

support workers was 31% and that by qualified staff,

69% (Table 3). The proportion of time patients spent withsupport workers was 27% and that with qualified staff,

73%. The mean time per contact for support workers

was 35.95 minutes compared to 75.76 minutes for quali-

fied staff. These results differed by team.

Factors associated with support worker

contribution to care

Univariate associations for outcomes of interest are pre-

sented in Table 4. Of the covariates considered for entry

(Box 1), a subset (TOMs, place in which care is being

delivered, team size, team integration and flexibility)

were found to be non-significant for all models. The

remaining covariates (the patient’s age, gender, EQ-5D atbaseline, level of dependence, professional referring,

usual living arrangements, intensity of intervention and

the percentage of support staff employed within their

care team) were generally found to be significant.

For ease of comparability, we present the same cova-

riates in the multivariate models for the two outcomes of

interest: (1) proportion of face-to-face patient time deliv-

ered by support workers (Tables 5 and 6) and (2) propor-tion of patient contacts delivered by support workers.

There were substantial missing data for two covariates

(EQ-5D at baseline, intensity of contact) both of which

were statistically significant. The approach taken was to



Table 1 Organisational characteristics of participating teams

Team ID Host

Location

(rural ⁄ urban)

WTE

clinical

staff*

No. diff.

type staff

Proportion of

WTE support

staff (%)‡Setting of care

provision

Patient

throughput (av

per annum)

A PCT Mixed 44.1 13 35 Client’s home 1800

B Social Services Urban 14.8 9 22 Client’s home 310

C PCT Urban 15.3 8 26 Client’s home 398

D PCT Rural 12.4 9 23 Client’s home 285

E PCT Rural 7.2 8 41 Client’s home 320

F Social Services Mixed 14.0 11 77 Resource Centre 183

G PCT Sub-urban 40.0 9 53 Client’s home 1650

J Acute Trust Urban 3.5 4 14 Accident and Emergency 777

L PCT Mixed 9.0 4 0 Client’s home 240

M PCT Urban 7.2 6 42 Client’s home 576

N PCT Urban 22.3 9 55 Client’s home 728

PA PCT 17.1 10 47 Client’s home 400

PB PCT Mixed 25.7 5 45 Community Hospital 166

Q Social Services Mixed 24.6 9 59 Client’s home 460

SA – – 24.3 8 24 Client’s home 365

SB PCT Sub-urban 50.7 11 37 Client’s home 2000

SG PCT Mixed 19.1 11 35 Client’s home 400

T PCT & Social Services Mixed 20.0 9 55 Client’s home 382

TA PCT & Social Services Mixed 19.1 10 46 Client’s home 267

U PCT Rural 6.0 4 50 Community hospital 533

PCT, Primary Care Trust; WTE, whole time equivalent.

*Excluding administrative staff.‡As a proportion of all clinical staff (not including administrative staff).

Table 2 Patient admission characteristics by team

Team ID n

Mean age

(SD)

Gender

(% female)

EQ-5D

(SD)

TOMs

Impairment

(SD)

TOMs

Activity

(SD)

TOMs

Participation

(SD)

TOMs

Wellbeing

(SD)

Modal level

of care need

(% patients)

A 313 80.2 (9.4) 193 (65) 0.4 (0.3) 3.3 (0.9) 3.2 (1.0) 3.2 (1.0) 3.7 (1.0) Level 4 (47)

B 85 83.8 (6.9) 64 (75) 0.5 (0.3) 3.1 (0.8) 3.3 (0.9) 3.2 (0.9) 3.6 (1.1) Level 3 (23)

C 18 72.3 (7.6) 12 (67) 0.3 (0.3) 3.3 (0.8) 2.9 (1.0) 2.7 (1.0) 3.7 (1.0) Level 5 (28)

D 53 80.5 (12.0) 32 (62) 0.4 (0.3) 2.8 (0.7) 3.0 (0.8) 3.3 (1.0) 3.7 (0.8) Level 4 (51)

E 69 78.6 (11.8) 46 (69) 0.3 (0.3) 3.1 (0.7) 3.0 (1.0) 3.4 (1.0) 3.8 (0.9) Level 3 (35)

F 52 82.4 (7.5) 40 (77) 0.4 (0.3) 3.2 (0.7) 3.3 (0.8) 3.7 (1.1) 3.9 (1.1) Level 4 (65)

G 173 78.0 (10.8) 113 (65) 0.3 (0.4) 3.3 (0.9) 3.2 (1.0) 3.3 (1.0) 3.9 (0.8) Level 3 (33)

J 81 81.5 (6.8) 65 (81) 0.5 (0.3) 3.2 (0.9) 3.2 (1.1) 3.5 (1.2) 4.1 (1.0) Level 1 (33)

L 30 77.8 (8.3) 18 (64) 0.5 (0.3) 2.5 (1.2) 3.3 (0.7) 3.2 (1.2) 3.6 (0.8) Level 6 (55)

M 98 82.0 (9.1) 58 (60) 0.4 (0.4) 3.3 (1.0) 3.1 (1.1) 3.5 (1.2) 3.9 (1.1) Level 4 (32)

N 100 80.5 (10.7) 66 (66) 0.5 (0.3) 3.1 (1.1) 2.9 (1.3) 3.2 (1.2) 3.7 (1.1) Level 1 (31)

PA 21 70.0 (12.6) 7 (39) 0.4 (0.3) 3.6 (1.1) 3.9 (0.6) 3.8 (0.8) 4.3 (0.5) Level 4 (64)

B 16 75.1 (12.0) 10 (71) 0.4 (0.3) 3.3 (0.9) 2.9 (0.9) 2.8 (0.7) 3.0 (0.8) Level 4 (81)

Q 46 80.3 (10.6) 28 (68) 0.6 (0.3) 2.9 (0.8) 2.9 (0.7) 2.7 (1.0) 3.5 (0.8) Level 4 (38)

SA 73 71.3 (14.2) 29 (43) 0.5 (0.3) 3.4 (1.0) 3.0 (1.1) 2.9 (1.0) 3.4 (1.0) Level 4 (41)

SB 225 82.9 (9.6) 144 (66) 0.3 (0.4) 3.1 (0.9) 3.2 (1.0) 2.8 (1.1) 3.6 (1.1) Level 4 (25)

SG 83 80.3 (6.1) 38 (68) 0.5 (0.3) 3.1 (0.8) 3.4 (0.9) 3.4 (1.2) 3.9 (0.9) Level 1 ⁄ Level

5 (26)

T 56 80.2 (11.2) 37 (76) 0.3 (0.4) 2.9 (1.0) 2.6 (1.1) 2.9 (1.2) 3.5 (1.1) Level 4 ⁄ Level

7 (28)

TA 241 76.5 (14.1) 158 (67) 0.4 (0.3) 3.1 (0.9) 3.1 (1.0) 3.2 (1.2) 3.6 (1.2) Level 3 (27)

U 49 82.4 (12.1) 32 (67) 0.2 (0.4) 2.5 (1.0) 2.6 (0.9) 2.7 (1.0) 3.3 (1.2) Level 7 (36)

All teams 1882 79.7 (11.0) 1190 (66) 0.4 (0.3) 3.1 (0.9) 3.1 (1.0) 3.2 (1.1) 3.7 (1.0) Level 4 (30)

TOMs, Therapy Outcome Measures.

A. Moran et al.


fit two models for each outcome: the first that includesthese terms (but uses less patients) (Table 5) and the sec-

ond that excludes them (an inferior model, but which

includes more patients) (Table 6). The resulting two

models allowed us to further investigate the relationship

between covariates and outcomes of interest with and

without the impact of patient dependency and intensity

of care.

Outcome (1): proportion of face-to-face patient timedelivered by support workersOn univariate analysis (Table 4), the most significant fac-

tors associated with a higher proportion of face-to-face

time with a support worker were increasing age (2.7%

increased support worker input for each increase in age

of 10 years) and being female (4.5% more input). None of

the dependency scores (TOMs or EQ-5D) was significantin this model.

However, clients requiring regular rehabilitation

(LOC 4) and intensive rehabilitation programmes (LOC

5) had around 12% more face-to-face time provided by a

support worker than clients who did not need any inter-

vention (LOC 0); required respite (LOC 2); or needed

treatment for an individual acute condition (LOC 6).

The source of referral produced surprising results,with patients who were referred from a community

hospital, and patients who had referred themselves hav-

ing 22% and 17% more face-to-face time with a support

worker respectively than those patients referred by a

general practitioner. The patient’s usual living arrange-

ments were also strongly associated with the proportion

of support worker input. Patients with more support (i.e.

living at home with a relative) received 10% less supportworker time than patients whose normal place of resi-

dence is a residential or nursing home. We included

‘usual living arrangements’ in the model on the assump-

tion that those patients who received their care at home,

and had a supporting relative, would require less sup-

port worker input. Patients who received their care in a

hospital received around 6% more of their care from sup-

port workers as compared with patients receiving care athome. More intensive models of care were also associ-

ated with higher support worker input.

Team characteristics (team size, measures of team

integration and role flexibility) were not associated with

proportion of support worker time. We included these

variables on the assumption that larger teams would be

more likely to have more support workers (Farndon &

Table 3 Proportion of face-to-face care delivered by support staff (by team)

Team ID

Total number of

face-to-face

contacts

Time per contact

support worker

(minutes)

Time per contact

qualified staff

(minutes)

Proportion of time

delivered by

support worker

(%)

Proportion of

contacts delivered

by support staff

(%)

Mean SD Mean SD Mean SD Mean SD

A 4908 43.4 28.1 81.0 17.7 29.1 22.1 34.7 25.8

B 2867 11.0 33.5 92.0 34.3 4.1 12.9 8.1 23.1

C 819 26.7 58.5 120.6 31.9 4.9 9.5 5.9 12.3

D 1960 34.0 24.6 66.6 20.8 37.1 28.9 41.6 30.9

E 1572 34.8 30.2 75.4 41.3 31.7 28.0 37.5 29.7

F 2749 28.6 22.2 55.0 26.0 49.5 34.2 56.3 35.4

G 1428 35.7 28.4 72.0 13.1 19.9 18.9 23.3 21.7

J 473 22.8 28.7 62.2 30.4 21.8 25.5 26.1 29.5

L* 343 1.9 6.8 72.9 18.5 0.7 2.1 1.3 4.6

M 1910 25.4 26.9 59.3 34.0 23.6 28.8 25.8 29.3

N 712 32.4 25.5 75.9 53.5 36.8 34.3 46.6 39.0

PA 469 35.4 40.1 100.9 34.1 24.6 31.3 27.7 33.2

PB – 88.1 12.2 88.7 12.5 40.7 3.4 40.9 5.0

Q 681 70.1 49.4 64.8 40.0 74.3 23.7 72.1 24.0

SA 2087 54.3 29.5 78.5 21.0 24.1 19.7 27.4 22.9

SB 5440 48.4 38.2 96.5 42.7 28.0 24.8 32.3 26.9

SG 65 28.7 18.3 52.3 16.6 17.9 14.3 21.7 14.7

T 1238 50.6 97.4 75.3 27.1 32.7 35.8 32.2 35.3

TA 5706 24.0 32.6 66.2 32.5 26.0 34.0 28.9 36.4

U 1356 32.7 32.8 80.9 27.9 28.0 29.9 30.2 31.2

All 36 783 36.0 37.7 75.8 32.9 27.0 28.1 30.8 30.4

*Team L reported that they had no support staff employed within their team (service proforma), but recorded the use of support staff

on occasion from other healthcare teams (staff activity data).



Nancarrow 2003) and that support worker utilisation

may be associated with better team integration (Nancar-

row 2004a, Moran et al. 2010) and more flexibility of staff

roles (Nancarrow 2003, Moran et al. 2010). Neither of

these relationships was shown to be true in this study.Not surprisingly, teams with a higher proportion of

support staff showed a slight increase in the proportion

of face-to-face time provided by support workers.

On multivariate analysis, model 1 (Table 5), incor-

porating the EQ-5D scores at baseline as a measure of

patient dependency and intensity of care, demonstrates

that patient dependency at admission had a significant

impact on the proportion of face-to-face time spentwith a support worker, with a 4.9% increase in face-

to-face support worker time for every 0.1 unit increase

in the EQ-5D (95% CI 0.3–9.6, P = 0.038). In other

Table 4 Univariate analysis for percentage of face-to-face patient time delivered by support workers (overall n = 1847)

Factor n

Unstandardised

regression

coefficient 95% CI P

Age (per 10 year increase) 1190 2.7 1.4, 4.0 <0.001

Gender (male as reference)* 4.5 1.7, 7.4 0.002

TOMs score on impairment at admission )0.3 )1.8, 1.2 0.701

TOMs score on activity at admission )0.5 )1.9, 0.9 0.476

TOMs score on participation at admission )0.1 )1.4, 1.1 0.835

TOMs score on wellbeing at admission 1.2 )0.1, 2.6 0.071

EQ-5D at admission (per 0.1 unit increase) 3.4 )0.9, 7.8 0.117

Level of dependence at baseline (level of care) <0.001

0 does not need any intervention 109

1 prevention ⁄ maintenance programme 248 2.7 )3.7, 9.1

2 convalescence ⁄ respite 43 )0.4 )10.3, 9.5

3 slow stream rehabilitation 315 8.3 2.0, 14.5

4 regular rehabilitation 488 12.0 6.0, 18.1

5 intensive rehabilitation 220 12.8 6.2, 19.4

6 specific treatment for acute condition 99 )1.1 )8.9, 6.8

7 medical care and rehabilitation 83 9.1 0.8, 17.3

8 rehabilitation for complex disabling condition 27 3.1 )8.5, 14.7

Who referred patient <0.001

General practitioner 330

Self ⁄ informal ⁄ carer ⁄ friend ⁄ family 22 17.0 5.1, 28.9

Community nurse 133 6.5 0.7, 12.3

Social worker 165 0.6 )5.3, 6.5

Allied health professional 384 5.2 0.7, 9.6

A&E ⁄ acute 135 4.2 )0.1, 8.6

Community hospital 103 22.2 15.6, 28.9

Other ICT ⁄ ward 13 10.2 )1.8, 22.3

Other 145 8.7 3.0, 14.3

Usual living arrangements 0.007

Own home, alone 804

Own home, with other(s) 529 )4.2 )7.2, )1.3

With relative 40 )8.1 )17.5, 1.3

Residential ⁄ nursing home 54 2.8 )4.5, 10.1

Sheltered housing 87 4.7 )1.5, 11.0

Other 11 )4.6 )21.0, 11.8

Where is care being delivered <0.001

Own ⁄ relative home 1201

Hospital 151 5.5 )9.0, 20.1

Community 305 2.0 )1.8, 5.7

Other 72 )22.1 )31.2, )13.1

Intensity (avg. no. minutes contact per day; log scale) 6.8 5.3, 8.4 <0.001

Team mean score for integration )0.1 )0.8, 0.6 0.770

Team mean score for role flexibility 0.2 )1.2, 1.5 0.806

Size of team 0.2 )0.3, 0.8 0.385

% of support staff in team 0.6 0.3, 0.9 <0.001

TOMs, Therapy Outcome Measures.

*On average, females have an additional 4.5% of their total contact time with support workers compared with males.

A. Moran et al.


words, a higher proportion of face-to-face support

worker time is provided to less-dependent patients.

Furthermore, intensity of care, as measured by average

minutes of contact per day, was found to be stronglyassociated with an increase in the proportion of face-

to-face time spent with a support worker (unstandar-

dised regression coefficient 6.8%, 95% CI 5.3–8.4,

P < 0.001). This suggests that greater support worker

involvement is associated with more intensive treat-

ment ⁄ care.

Model 1 (Table 5) also supports our assumption that

patients who normally live at home with a relative aresignificantly less likely to receive support worker input

than any other group of patients, but particularly those

who usually live in a nursing or residential home. The

relationship between source of referral and proportion of

face-to-face time with a support worker, however, does

not vary substantially between the two models (Tables 5

and 6), suggesting that there is little or no relationship

between the source of referral and patient dependency.

In both models, patients referred to the service by them-

selves ⁄ family, or from a community hospital have

between 10% and 20% more support worker input thanthose referred by a GP. What we would expect to see,

however, would be patients referred by hospital wards

are more likely to be more dependent, and therefore,

have less input from a support worker and those

referred by a community hospital are more likely to be

referred home for enablement care, which is often deliv-

ered by support workers (Nancarrow et al. 2009, Moran

et al. 2010).When the effects of patient dependency (as mea-

sured by EQ-5D) and intensity of care on the propor-

tion of face-to-face time spent with a support worker

are removed from the model (Model 2, Table 6), the

effect of age increases from 1.1% to 2% increased

support worker input for each increase in age of

10 years.

Table 5 Multivariate model for percentage of face-to-face patient time delivered by support workers [Model 1: incorporating EQ-5D at

baseline and intensity of delivery (N = 1135)]

Factor

Unstandardised

regression coefficient 95% CI P

Age (per 10 year increase) 1.1 )0.4, 2.6 0.163

Gender (male as reference)* 5.1 1.9, 8.4 0.002

EQ-5D at admission (per 0.1 unit increase) 4.9 0.3, 9.6 0.038

Level of dependence at baseline (level of care)

0 does not need any intervention 0.079

1 prevention ⁄ maintenance programme )7.0 )19.4, 5.4

2 convalescence ⁄ respite )0.1 )15.8, 15.7

3 slow stream rehabilitation )5.4 )17.7, 6.8

4 regular rehabilitation )0.6 )12.7, 11.4

5 intensive rehabilitation )0.1 )12.5, 12.3

6 specific treatment for acute condition )8.6 )21.8, 4.6

7 medical care and rehabilitation )3.0 )16.7, 10.7

8 rehabilitation for complex disabling condition )1.3 )18.9, 16.2

Who referred patient

General Practitioner <0.001

Self ⁄ informal ⁄ carer ⁄ friend ⁄ family 11.0 )2.6, 24.5

Community nurse 4.8 )2.0, 11.6

Social worker )6.2 )12.4, 0.0

Allied health professional )1.1 )5.9, 3.7

A&E ⁄ acute )1.4 )6.1, 3.3

Community hospital 19.3 12.4, 26.2

Other ICT ⁄ ward 4.6 )7.8, 16.9

Other 4.7 )1.6, 11.0

Usual living arrangements

Own home, alone 0.036

Own home, with other(s) )1.1 )4.6, 2.3

With relative )11.8 )22.1, )1.5

Residential ⁄ nursing home 11.9 2.0, 21.8

Sheltered housing 1.5 )5.5, 8.5

Other 0.7 )18.3, 19.8

Intensity (avg. no. minutes contact per day; log scale) 7.4 6.0, 8.8 <0.001





Outcome (2): proportion of patient contacts delivered bysupport workersThe pattern of the relationships described above remains

the same when the percentage of face-to-face contacts isused as the independent variable: age (unstandardised

regression coefficient 2.4%, 95% CI 1.0–3.9, P = 0.001),

gender (unstandardised regression coefficient 2.9%, 95%

CI 0.4 to 6.1, P = 0.081) and % support staff in the team

(unstandardised regression coefficient 0.4%, 95% CI 0.3–

0.5, P < 0.001).

Discussion

This research demonstrates that although support work-

ers within CBRS do not undertake the greatest propor-

tion of the direct patient care workload compared with

qualified practitioners, the amount of care they provide

is associated with a number of patient- and team-level

factors. In summary, the results from this study demon-

strate that a higher proportion of input from support

workers, as measured by time and total face-to-face con-

tacts, is associated with the following patient and team

characteristics: decreasing patient dependency at admis-

sion, older age, being female, patients who normally livein a residential or nursing home, self-referral to the ser-

vice, referral by a community hospital, receiving a more

intensive model of care and having a higher proportion

of support staff in the team.

Research has demonstrated that there is little differ-

ence in the types of tasks undertaken by any level ⁄type of staff in acute hospitals in the UK (Jenkins-

Clarke & Carr-Hill 2003). The results of this study,however, indicate that CBRS may be better at differen-

tiating workload between staff types. Associations

were found between the proportion of support work-

ers in the team and the proportion of care delivered

by support staff, to the effect that a 1% increase in

proportion of support staff in the team was associated

with a 0.4% increase in support worker input with

patients (95% CI 0.3–05, P < 0.001).

Table 6 Multivariate model outcome for percentage of face-to-face patient time delivered by support workers [Model 2: excluding

EQ-5D at baseline and intensity of delivery (N = 1463)]

Factor

Unstandardised

regression coefficient 95% CI P

Age (per 10 year increase) 2.0 0.6, 3.4 0.006

Gender (male as reference)* 2.9 )0.2, 6.1 0.069

Level of dependence at baseline (level of care)

0 does not need any intervention <0.001

1 prevention ⁄ maintenance programme )0.4 )7.3, 6.5

2 convalescence ⁄ respite )0.3 )10.8, 10.2

3 slow stream rehabilitation 3.7 )3.1, 10.5

4 regular rehabilitation 8.8 2.2, 15.3

5 intensive rehabilitation 9.3 2.3, 16.3

6 specific treatment for acute condition )2.6 )10.8, 5.6

7 medical care and rehabilitation 8.7 0.1, 17.3

8 rehabilitation for complex disabling condition 1.9 )11.2, 14.9

Who referred patient

General Practitioner <0.001

Self ⁄ informal ⁄ carer ⁄ friend ⁄ family 14.9 1.5, 28.3

Community nurse 6.9 0.6, 13.2

Social worker )4.0 )9.9, 1.9

Allied health professional 4.1 )0.4, 8.7

A&E ⁄ acute 4.8 0.4, 9.1

Community hospital 23.8 16.7, 30.8

Other ICT ⁄ ward 10.1 )2.5, 22.7

Other 7.1 1.1, 13.1

Usual living arrangements

Own home, alone 0.257

Own home, with other(s) )2.3 )5.6, 1.0

With relative )9.9 )20.1, 0.3

Residential ⁄ nursing home 3.2 )4.4, 10.8

Sheltered housing 0.9 )5.8, 7.5

Other )0.3 )17.6, 17.1



A. Moran et al.


This suggests that as support worker numbers in theteam grow as a proportion of all staff members, so does

the proportion of care they deliver. Although a stronger

relationship was expected, as implied in the policy and

the evidence base, the association found may be partially

explained by the small amount of data available to deter-

mine the proportion of support staff at team level

(n = 20). The observed association, however, is further

supported by the finding that as the intensity of care pro-vided increased, so too did the proportion of support

worker input.

Another policy and service assumption that has

been partially confirmed by this research is the expec-

tation that support workers are utilised to ‘free-up’

qualified staff time (Russell & Kanny 1998, Steele &

Wright 2001, Mackey 2004, Taylor & Birch 2004). That

is, if qualified professionals focus their expertise andtime on more dependent patients with severe or com-

plex conditions (Anderson 1997, Saunders 1998,

Thornley 2000, Bowman et al. 2003, Leigh 2003, Reid

2004, Taylor & Birch 2004, Keeney et al. 2005), then

support workers should be spending a greater propor-

tion of their time with less impaired clients. The

results of this study demonstrate that in general this

was the case within CBRS with a higher proportion offace-to-face support worker time found to be directed

towards less-dependent patients. Once again, the

strength of the association is small, perhaps reflecting

the myriad external pressures that can and do impact

the organisation and type of care delivered by these

teams (McClimens et al. 2010).

The relationship between living arrangements and

support worker input is interesting. It suggests that liv-ing with a relative substantially reduces the likelihood of

a patient being seen by a support worker, indicating that

support workers are fulfilling some of the role normally

provided by family members. The trend for support

workers to play a carer role when providing health and

social care has been documented in other research (Bow-

man et al. 2003, Stanmore & Waterman 2007, Moran et al.2010). This relationship may also, in part, explain theincreased use of support workers by women. Older

women may be more likely to be living alone than older

men (Soule et al. 2005), thus increasing their need for

support.

As Nancarrow (2004a) and McClimens et al. (2010)

demonstrate, staff within CBRS perceive there to be a

clear division of labour between qualified staff and sup-

port workers, with support workers carrying out themajority of care provision and qualified practitioners

undertaking assessment, triage and care planning. Fur-

thermore, there is an overall acceptance that the provi-

sion of direct care is no longer effective use of qualified

practitioner time. Although the staff activity data suggest

that this division of labour is not as definite as that whichis perceived by CBRS staff and policy-makers, this

research goes some way to support these perceptions.

The findings also suggest that there may be further

potential for both groups of workers to be used more

systematically and effectively. For example, there is

potential for support workers to be deployed more sys-

tematically in patient care, possibly according to patient

severity, and for qualified staff to focus more on their‘expert’ role for more complex cases. These suggestions

reflect those reported in other workforce research. Sib-

bald, for example, has recently suggested that efficiency

gains are possible if general practitioners discontinue the

services that practice nurses provide and focus on the

tasks only doctors can perform (Sibbald 2008). Further-

more, analysis of patient outcomes as part of the longitu-

dinal component of this study demonstrated that agreater proportion of support worker input was associ-

ated with improved patient outcomes (Nancarrow et al.2010). This suggests that more focused utilisation of the

workforce according to patient admission characteristics

may have the potential to further influence the gains

observed in patient outcomes.

Study limitations

Workforce research is not straightforward. In an ideal

world, it would be possible to control for all external and

internal variables and therefore have a preferential skill-

mix with staff working optimally. These teams, however,

do not operate in a vacuum. There are many variablessuch as poor supporting services or staffing shortages

that influence their workforce efficiency (Nancarrow

2004b, McClimens et al. 2010). These variables may there-

fore account for the lack strength found in some of the

relationships explored between patient- and team-level

factors and support worker utilisation.

A key limitation to this study is the extent to which

the teams sampled and presented in this research arerepresentative of other community rehabilitation teams

in England and the heterogeneity of team composition,

structure and function. To overcome this limitation, the

sample of 27 teams was drawn from a much wider sur-

vey, which was sent to over 500 teams nationally (Nanc-

arrow et al. 2009a). The classification of and therefore

inclusion criteria for participating teams were based on a

comprehensive thematic analysis of policy documentsand peer-reviewed journals that sought to ‘define’ CBRS

(Nancarrow et al. 2009b). However, as has been reported

elsewhere, it is difficult to clearly define a community-

based rehabilitation team (Enderby & Stevenson 2000;

Enderby & Wade 2001) and the variability in teams, staff-

ing ratios and patient profiles within CBRS that has been

observed in this study and others (Enderby & Wade



2001; Nancarrow et al. 2009) reduces the strength of theconclusions we can draw. This reinforces the notion that

this is a complex area to research, and the heterogeneity

of community rehabilitation teams and lack of systematic

data collection to date mean that few studies attempt

comparisons at this level. To our knowledge, this is the

only study that has attempted to examine the relation-

ship between multidisciplinary staffing ratios in commu-

nity rehabilitation teams and patient profiles. Clearly,more research is needed in this area to help inform the

debate about the most effective, and cost-efficient way to

staff CBRS and the best way to measure skill-mix

changes.

This study has drawn on three main sources of data:

patient-level data, staff-level data and team-level data.

To undertake comparisons between the variables at a

team level has meant aggregating the findings fromsome of these variables, reducing the numbers of obser-

vations to 20, thus reducing the strength of the study to

draw conclusions at this level.

We have analysed and drawn conclusions on data

based on professional title or the difference between

‘support worker’ and ‘qualified professional’ rather than

on the specific roles carried out. Given the large

variations in the roles ascribed to each category, thesetitles are unlikely to be true reflections of the complexity

of the work performed by these practitioners. However,

recent evidence demonstrates that despite the heteroge-

neity of roles performed by support workers, there are

key generic features common to all support roles regard-

less of context (Moran et al. 2010).

We used EQ-5D in the multivariate analyses, as it is a

unidimensional measure of dependency, as opposed toTOMs, which has four domains. There is therefore a

chance that the choice of the EQ-5D over TOMs may

have limited the conclusions drawn from the analyses.

The LOC is also a measure of dependency, and inclusion

of this measure in the multivariate analyses reinforced

the relationship between dependency and support

worker input (although did not necessarily define the

direction of that dependency).The other significant limitation is that which is inher-

ent within an observational study, namely, that the vari-

ous relationships do not imply causality, and nor do

they suggest the direction of any causality. So, whilst

plausible explanations are possible that match up with

theory, rationales for policy and ⁄or intuition with each of

the identified relationships, these are best tested in a con-

trolled evaluation.

Conclusion

The ability and effectiveness of health and social care

services to delineate the workload carried out by

qualified and support workers is an important policyand service imperative. The findings of this research

demonstrate that rehabilitation teams for older people

based in the community are utilising their sup-

port:qualified skill-mix according to a number of

patient and team characteristics. In particular, this

research has provided evidence to suggest that sup-

port workers deliver a high proportion of care to

older people in the community and that they aremore likely to be involved in the care of older, female,

less-dependent patients. These findings suggest that

the targeted utilisation of support workers within a

support:qualified skill mix may enable a more effec-

tive use of qualified practitioner time and as such fur-

ther gains in workforce efficiency may be achieved.

Despite the findings of this research, there remains a

need for greater acknowledgement of the complexitiesthat influence workforce efficiency to provide more accu-

rate measures of the impact of skill-mix changes. In par-

ticular, that there are innumerable variables that will

have varying degrees of influence on how effective sup-

port workers are when added to the skill-mix and conse-

quently, the impact these variables have on the desired

effect they are intended to bring about.

Acknowledgement

This study was funded by National Institute for Health

Research Service Delivery and Organisation (NIHR SDO).

References

Anderson L. (1997) The introduction of generic workersinto the ward team: an exploratory study [erratumappears in J Nurs Manag 1997 Sep;5(5):319]. Journal ofNursing Management 5, 69–75.

Barton P., Bryan S., Glasby J. et al. (2005) A National Evalua-tion of the Costs and Outcomes of Intermediate Care for OlderPeople. Policy Research Programme, London, The Univer-sity of Birmingham and The University of Leicester.

Bowman S., Bray K., Leaver G., Pilcher T., Plowright C. &Stewart L. (2003) Health care assistants’ role, functionand development: results of a national survey. Nursing inCritical Care 8, 141–148.

Buchan J. & Dal Poz M.R. (2002) Skill mix in the healthcare workforce: reviewing the evidence. Bulletin of theWorld Health Organization 80, 575–580.

Department of Health (2000a) A Health Service of All the Tal-ents: Developing the NHS Workforce. Consultation Documenton the Review of Workforce Planning. HMSO, London.

Department of Health (2000b) Meeting the Challenge: AStrategy for the Allied Health Professions. HMSO, London.

Department of Health (2001) Working Together, LearningTogether: A Framework for Lifelong Learning for the NHS.HMSO, London.

Department of Health (2004) Social Care Workforce: Develop-ing a New National Agenda. HMSO, London.

A. Moran et al.


Dolon P. (1997) Modeling valuations for EuroQol healthstates. Medical Care 35, 1095–1108.

Enderby P. & Stevenson J. (2000) What is intermediate care?Looking at needs. Managing Community Care 8, 35–40.

Enderby P. & Wade D. (2001) Community rehabilitationin the United Kingdom. Clinical Rehabilitation 15, 577–581.

Enderby P., John A. & Petherham B. (2006) Therapy Out-come Measures for Rehabilitation Professionals. John Wileyand Sons Ltd., Chichester.

Farndon L. & Nancarrow S. (2003) Employment and careerdevelopment opportunities for podiatrists and foot care assis-tants in the NHS. British Journal of Podiatry 6 (4), 103–108.

Hart E., Lymbery M. & Gladman J. (2005) Away fromhome: an ethnographic study of transitional rehabilitai-ton scheme for older people in the UK. Social Science &Medicine 60, 1241–1250.

Jenkins-Clarke S. & Carr-Hill R. (2003) Workforce andWorkload: are nursing resources used effectively? Nurs-ing Times Research 8, 238–248.

Keeney S., Hasson F., Mckenna H. & Gillen P. (2005) Nurses’,midwives’ and patients’ perceptions of trained health careassistants. Journal of Advanced Nursing 50, 345–355.

Knight K., Lamer S. & Waters K. (2004) Evaluation of therole of the rehabilitation assistant. International Journal ofTherapy and Rehabilitation 11 (7), 311–317.

Leigh J. (2003) How a new health care assistant role cansupport junior doctors. Nursing Times 99, 32–34.

Mackey H. (2004) An extended role for support workers:the views of occupational therapists. International Journalof Therapy and Rehabilitation 11, 259–266.

McClimens A., Nancarrow S., Moran A., Enderby P. &Mitchell C. (2010) ‘Riding the bumpy seas’: the impact ofthe Knowledge Skills Framework component of theAgenda for Change initiative on staff in intermediatecare settings. Journal of Interprofessional Care 24, 70–79.

Moran A., Enderby P. & Nancarrow S. (2010) Defining andidentifying common elements of and contextual influ-ences on the roles of support workers in health andsocial care: a thematic analysis of the literature. Journal ofEvaluation in Clinical Practice, doi:10.1111/j.1365-2753.2010.01505.x.

Nancarrow S. (2003) Promoting flexibility in the intermedi-ate care workforce. Journal of Integrated Care 11, 25–30.

Nancarrow S. (2004a) Dynamic role boundaries in intermedi-ate care services. Journal of Interprofessional Care 18, 141–151.

Nancarrow S. (2004b) Improving intermediate care: givingpractitioners a voice. Journal of Integrated Care 12, 33–41.

Nancarrow S.A., Moran A., Freeman J., Enderby P., DixonS., Parker S. & Bradburn M. (2009) Looking inside theblack box of community rehabilitation and intermediatecare teams in the United Kingdom: an audit of serviceand staffing. Quality in Primary Care 17, 323–333.

Nancarrow S.A., Moran A.M. & Parker S.G. (2009a) Under-standing service context: development of a service proforma to describe and measure elderly peoples’ commu-nity and intermediate care services. Health & Social Carein the Community 17 (5), 434–446.

Nancarrow S., Moran A., Freeman J., Enderby P., Dixon S.& Parker S. (2009b) Looking inside the black box of com-munity rehabilitation and intermediate care teams in theUnited Kingdom: an audit of service and staffing config-uration. Quality in Primary Care 17 (5), 323–333.

Nancarrow S., Enderby P., Moran A. et al. (2010) The Rela-tionship Between Workforce Flexibility and the Costs and Out-comes of Older People’s Services. National Institute forHealth Research, Southampton.

Patient Information Advisory Group (2006) Informationabout Patients. An Introduction to the Patient InformationAdvisory Group for Health Professionals and Researchers.HMSO, London.

Reid C. (2004) Developing a tissue viability nursing assis-tant role. Nursing Standard 18, 68–72.

Rolfe G., Jackson N., Gardner L., Jasper M. & Gale A.(1999) Developing the role of the generic healthcare sup-port worker: phase 1 of an action research study. Interna-tional Journal of Nursing Studies 36, 323–334.

Russell K.V. & Kanny E.M. (1998) Use of aides in occupa-tional therapy practice. American Journal of OccupationalTherapy 52, 118–124.

Saks M. & Allsop J. (2007) Social policy, professional regu-lation and health support work in the United Kingdom.Social Policy and Society 6, 165–177.

Saunders L. (1998) Managing delegation, a field study of asystematic approach to delegation in out-patient physio-therapy. Physiotherapy 84, 547–555.

Shrive F.M., Stuart H., Quan H. & Ghali W.A. (2006) Deal-ing with missing data in a multi-question depressionscale: a comparison of imputation methods. BMC MedicalResearch Methodology 6, 57.

Sibbald B. (2008) Should primary care be nurse led? Yes.British Medical Journal, doi: 10.1136/bmj.39661.707083.59.

Stanmore E. & Waterman H. (2007) Crossing professionaland organizational boundaries: the implementation ofgeneric rehabilitation assistants within three organizationsin the northwest of England. Disability & Rehabilitation29, 751–759.

Stanmore E., Ormrod S. & Waterman H. (2005) New rolesin rehabilitation – the implications for nurses and otherprofessionals. Journal of Evaluation in Clinical Practice 12,656–664.

Steele V. & Wright R. (2001) Cutting it fine. Health ServiceJournal 111, 28.

Stevenson D. (2000) Rehabilitation outreach: a groundedtheory study. British Journal of Therapy & Rehabilitation 7,112–115.

Taylor L. & Birch J. (2004) Chemotherapy supportworker role: a one-year pilot. Cancer Nursing Practice 3,23–27.

The Euroqual Group (1990) EuroQol: a new facility for themeasurement of health related quality of life. Health Pol-icy 16, 199–208.

Thornley C. (2000) A question of competence? Re-evaluatingthe roles of the nursing auxiliary and health care assistantin the NHS. Journal of Clinical Nursing 9, 451–458.



are we using support workers effectively? the relationship between patient and team characteristics...

Documents