Research

Jackie Sturt, Saima Ali, Wendy Robertson, David Metcalfe, Amy Grove, Claire Bourne
and Chris Bridle

Neurolinguistic programming:
a systematic review of the effects on health outcomes

INTRODUCTION to the development of rapport, to facilitate

Neurolingistic programming (NLP) is modelling, elicit well formed outcomes and
an emerging technology within health use anchoring (or conditioning) techniques.
Abstract care attracting interest and investment, NLP training is informally regulated in the
Background particularly within primary care. NLP is a UK, through the Association of NLP (ANLP)3
Neurolinguistic programming (NLP) in health communication framework using techniques and internationally through the International
care has captured the interest of doctors, to understand and facilitate change in NLP Trainer’s Association (INLPTA)4 at three
healthcare professionals, and managers.
thinking and behaviour. Early study of NLP levels of diploma, practitioner, and master
Aim was of a scholarly nature and promoted practitioner, based on the number of hours
To evaluate the effects of NLP on health- NLP as a psychotherapeutic technique, of study and practice.3,4 While NLP training
related outcomes.
although publication of commercial works1,2 organisations and practitioner registers
Design and setting in the 1980s signalled a move between the are internationally widespread with NLP
Systematic review of experimental studies. academic and commercial worlds. While training opportunities for business use,
Method there is no agreed definition of NLP, different personal development, and health visible in
The following data sources were searched: formulations share (or practitioners accept) many European countries, US, Canada, and
MEDLINE®, PsycINFO, ASSIA, AMED, a set of core propositions. In particular, NLP Australia,3–6 the targeting of medical and
CINAHL®, Web of Knowledge, CENTRAL, NLP
specialist databases, reference lists, review
proposes that our internal representations healthcare practitioners for such training by
articles, and NLP professional associations, of the world show a bias for a particular NLP organisations in the form of seminars,
training providers, and research groups. sensory modality (visual, auditory, workshops, and literature appears to be
Results kinaesthetic, olfactory or gustatory), and presently focused on the UK.7–10
Searches revealed 1459 titles from which that a person’s dominant modality, or This targeted interest by the NLP
10 experimental studies were included. Five preferred representational system (PRS), community in medical and healthcare
studies were randomised controlled trials is signalled through various behavioural professionals led the authors to make a
(RCTs) and five were pre-post studies. Targeted
health conditions were anxiety disorders,
indices, particularly verbal expression and UK Freedom of Information (FOI) request
weight maintenance, morning sickness, eye movement. A visual person, for example, to NHS organisations to identify spending
substance misuse, and claustrophobia during may say ‘I see what you mean’ whereas an on NLP training or services over a 3-year
MRI scanning. NLP interventions were mainly auditory thinker may say ‘I hear what you period. Information was requested on
delivered across 4–20 sessions although
three were single session. Eighteen outcomes
say’. The central tenet, or hypothesis of NLP the purpose of any training (for example,
were reported and the RCT sample sizes is that communication will be more effective, personal development, management
ranged from 22 to 106. Four RCTs reported no or persuasive, if it is tailored to match the training, clinical service provision), which
significant between group differences with the PRS of the target person. NLP practitioners staff were trained and whether any
fifth finding in favour of the NLP arm (F = 8.114,
P<0.001). Three RCTs and five pre-post studies
use the individual’s PRS as a foundation associated evaluations or audits had been
reported within group improvements. Risk of
bias across all studies was high or uncertain.
J Sturt, PhD, BA(Hons), RGN, RMN, professor Address for correspondence
Conclusion
of behavioural medicine in nursing, Florence Jackie Sturt, Florence Nightingale School of
There is little evidence that NLP interventions
Nightingale School of Nursing and Midwifery, Nursing and Midwifery, King’s College London.
improve health-related outcomes. This
King’s College London, London. W Robertson, James Clerk Maxwell Building, 57 Waterloo Road,
conclusion reflects the limited quantity and
BSc, MPH, MSc, MPhil, PhD, FFPH, associate London SE1 8WA, UK.
quality of NLP research, rather than robust
evidence of no effect. There is currently professor; D Metcalfe, BSc(Hons), MBChB, E-mail: [email protected]
insufficient evidence to support the allocation LLB(Hons), NIHR academic clinical fellow;
Submitted: 7 September 2011; Editor’s response:
of NHS resources to NLP activities outside of NIHR academic clinical fellow; A Grove, BSc,
MSc, research fellow; C Bridle, PhD, CPsychol, 26 October 2011; final acceptance: 8 March 2012.
research purposes.
principal research fellow, Warwick Medical School, ©British Journal of General Practice
Keywords University of Warwick, Coventry. S Ali, BSc, MSc, This is the full-length article (published online
experimental designs; neurolinguistic PhD, research fellow, University of Southampton, 29 Oct 2012) of an abridged version published in
programming; primary care; review, Southampton. C Bourne, BSc (Hons), MSc, senior print. Cite this article as: Br J Gen Pract 2012;
systematic; treatment effectiveness. research fellow, Coventry University, Coventry. DOI: 10.3399/bjgp12X658287

e757 British Journal of General Practice, November 2012

 effectiveness of NLP on health-related
outcomes.
How this fits in
Neurolinguistic programming (NLP) METHOD
is a collection of communication and Studies were eligible for inclusion if they
behaviour change techniques used within reported primary research on the effects
the NHS for both clinical and managerial of NLP on any health-related outcomes in
purposes and has a reputation in the
all clinical populations. Studies without a
business and entertainment industry as
a method for influencing people. NLP is
quantitative evaluation of the effect of NLP,
promoted to health professionals as a single case (n = 1) studies, and those in which
therapeutic and managerial intervention. a single NLP technique was evaluated were
Limited experimental research has excluded. Language eligibility was restricted
been undertaken into the use of NLP to English.
to influence health outcomes and there MEDLINE® Ovid version (1950 to
is little evidence that NLP interventions 20/02/12), PsycINFO (Earliest to 20/02/12),
improve health outcomes based on poor
Applied Social Science Index and Abstracts
quality studies across heterogeneous
conditions and populations. The allocation (ASSIA) (Earliest to 20/02/12), Allied and
of NHS resources to support NLP Complementary Medicine Database
activities should be confined to research (AMED) (1985 to 20/02/12), CINAHL® (1981
investigations. to 20/02/12) , Web of Knowledge (Earliest to
20/02/12) and CENTRAL (Earliest to 20/02/12)
were searched. The following keywords
‘neurolinguistic/neuro-linguistic and neuro
undertaken. This request was sent in June linguistic programming’ were combined
2009 to all 143 primary care trusts, 73 using the ‘OR’ Boolean operator together
mental health trusts, 166 hospital trusts, with the MeSH heading ‘neurolinguistic
12 ambulance trusts, 10 care trusts, and 10 programming’ (available for MEDLINE
strategic health authorities. A total of 326 only). The specialist NLP databases at the
(79%) NHS organisations responded to the Universities of Bielefeld and Surrey (to
request and the unpublished data revealed 20/02/12), and the European Association for
an NHS monetary spend of £802 468 on Neuro-linguistic Psychotherapy (to 20/02/12)
NLP-related activity. Over 700 NHS staff were also searched in their entirety, and
undertook NLP training during the time NLP associations, research groups, and
period with the majority (75%) being in social network forums, were contacted
administrative/managerial roles. Clinical for additional research. Reference lists
staff included counsellors and clinical were screened for additional citations. A
psychologists. A conservative estimate of single reviewer initially screened all search
1-day training per person was determined results by title and those deemed potentially
at a modest daily salary rate of £150 per relevant were assessed against the eligibility
person indicates an estimated training cost criteria by two reviewers independently, with
of £105 000. For five trusts reporting that discrepancies resolved through discussion.
they had developed NLP-based services, Full-text papers of included studies were
the majority was spent on weight-loss assessed against the eligibility criteria by two
counselling (£200 000) and this was a reviewers independently and discrepancies
research study. Other spend areas included were resolved through discussion or
counselling skills (£190), substance misuse referral to a third reviewer.
counselling (£90) and smoking cessation A data extraction template was
services (£450). While this spend was found developed, pilot-tested on two papers by
to be modest, the FOI request established three reviewers, and modified as necessary.
that it was widespread. Two reviewers independently extracted
NLP’s position outside mainstream data from each study, including: publication
academia has meant that while the details (authors, year, and country),
evidence base for psychological intervention participant characteristics, intervention
in both physical and mental health has details, outcome measures, risk of bias, and
strengthened,11–14 parallel evidence in study findings. Risk of bias assessment for
relation to NLP has been less evident and the randomised controlled trials (RCTs) was
has attracted academic criticism.15,16 No undertaken with reference to the Cochrane
systematic review of the NLP literature Handbook.17 The RCTs were assessed
has been undertaken applying Cochrane against the four risk domains of sequence
methods.17 The aim of this study was to generation, allocation concealment, blinding
conduct a systematic literature review of outcome assessors, and incomplete
and appraise the available evidence for outcome data. The risk of bias was assessed

British Journal of General Practice, November 2012 e758

 Table 1. Details of included studies
Author, country Study design Population and NLP Assessed outcomes
and health issue and setting participants Intervention details and measures Results
Randomised Controlled Trials
Krugman et al 19 Design: RCT Participants: university NLP interventionist training: Public speech anxiety. Between group differences; no statistical difference
US Setting: University undergraduate students three graduate clinical/counselling Personal Report of Confidence between groups (results data not reported). Within
Speech anxiety Assessmen: Baseline and Recruitment: response to psychologists with additional as a Speaker Scale. Paul’s group differences (pre-post tests): t values
immediately post-treatment advertisement for a programme 4 months of NLP training. Modified Behaviour (pre-post tests): t values attained statistical
Groups: arm 1) NLP single to alleviate anxiety in public Training fidelity checks employed Checklist. Observed global significance (P<0.05) showing reduction in all
session; arm 2) self-controlled speaking situations. Study NLP intervention: phobia rating of speech anxiety measures of speech anxiety (including fear expectancy
desensitisation; arm 3) waiting baseline, n = 55, 28 male/27 female. intervention from ‘Frogs into and fear survey) in all three arms.
list control Numbers randomised, analysed Princes’:2 Kinaesthetic
and completed not reported Anchoring techniques, visualisation
Single session

e759 British Journal of General Practice, November 2012

de Miranda et al 20 Design: RCT. Setting: day Participants: mother and infant NLP interventionist training: Child development (Bayley Between Groups: Non-significant trend towards
Brazil care centre. Assessment: pairs. Mothers’ age unreported, not described. NLP intervention: scale). Home environment improvement in the HOME environment (P = 0.09).
Maternal emotional baseline and 9 months child age 18–36 months, sex arm 1) not described beyond variation (Home Observation Variations in child mental development (OR 1.21,
disturbance and post-intervention. Groups: unreported. Recruitment: not ‘NLP’, 15 sessions over 1 year for Measurement of the 95% CI = 0.0 to 23.08, P = 0.669). Maternal mental
child emotional arm 1) NLP; arm 2) control, described. Randomised n = 45 Environment [HOME] health: P = 0.26.
development no further description (23 NLP/22 control). questionnaire). Maternal
Completed n = 37. Analysed: Mental Health (Self
children n = 27, mothers n = 37 Report Questionnaire)
(10 NLP/27 control)
Stipancic et al21 Design: RCT waiting list Participants: self-referred for NLP interventionist training: Baseline/screening Between groups: ANOVA test found NLP arm
Croatia controlled trial alternately psychological difficulties to ‘reduced Seven psychotherapists trained structured clinical interviews resulted in QOL improvement and decreased clinical
Psychological allocated. Setting: private rate’ psychotherapy: 79% female to NLP master practitioner level. for DSM-IV Personality symptoms (F = 8.114, P = 0.000). Within group findings
difficulties psychotherapy practice. <21 years = 9%; 21–40 years = 59% NLP intervention: individual Disorders. Croatian scale of at 5 months found improvement in intervention
Assessments: arm 1) >40 years = 31%. Married 24%; neurolinguistic psychotherapy Quality of Life compared to baseline (F=3.672, P = 0.019. A small
baseline, post-treatment and employed 56%; college educated or (NLPt). Weekly x 60 min sessions. significant interaction was found between number
5 months post-treatment. higher 100%. Recruitment: multiple Mean n = 20 (range 5–65). of sessions and size of improvement.
Arm 2) baseline, 3 months. methods. Alternately allocated
Groups: arm 1) NLP n = 106 (54 int/52 wait list)
psychotherapy, arm 2) Number completed not reported.
Wait list control Analysed n = 54 in int group.
Control group analyses
not reported.
Sorensen et al 22 Design: RCT. Setting: weight Participants: overweright or NLP interventionist training: Fasting weight in light clothes Between groups: during the 5 months of treatment,
Denmark loss clinic. Assessment: obese adults aged experienced NLP practitioner to nearest 0.1kg on digital there was no significant difference observed between
Weight baseline pre randomisation, 25–55 years. Recruitment: (certified by Danish NLP institute). scales groups for additional weight lost: NLP: –1.8kg versus
maintenance post intervention and at 2 At weight loss clinic following NLP intervention: NLP course cookery course: –0.2kg (ANCOVA, NS). After 3 years
and 3 years. Groups: arm 1) >8% weight loss during arm 1) 10 sessions over 5 months, 57% in the NLP group and 50% in the cookery group
NLP therapy, arm 2) a course 12-week programme. arm 2) 10 sessions over 5 months had maintained a part of their initial weight loss. There
in gourmet cookery Number in study: randomised was no significant difference observed between
n =  48 (23 to NLP, 25 cookery); groups (ANCOVA)
completed n = 41 (17 NLP,
24 cookery); analysed at 3 years 
n =  34 (16 NLP, 18 cookery)
 Simpson and Design: equivalence Participants: adults meeting NLP interventionist: trained Hospital Anxiety and Between groups: 4 week follow-up — no between
Dryden 23 randomised trial. Setting not DSM-IV criteria for panic disorder hypnotherapist (registered with Depression Scale (HADS) group differences: HADS depression: F = 0.106,
UK reported. Assessments: age range 23–65 years. Recruited UK Council for Psychotherapists Agoraphobic Cognitions P = 0.749 HADS anxiety: F = 0.003, P = 0.96;
Panic Disorder screening, baseline post via media advert. Randomised  as a hypno-psychotherapist) Questionnaire (ACQ) ACQ: F = 0.374, P = 0.549; PASQ: F = 0.659, P = 0.429;
randomisation, immediately n = 22, completed n = 18, NLP Intervention: VKD also known Panic Attack Symptoms GPR: F = 3.586, P = 0.076. There was a greater change
at intervention completion and (12 females/6 males) nine in each as Fast Phobia technique. Four Questionnaire (PASQ) in the pre-post scores for the VKD arm as follows pre-
4 weeks post completion. arm. Mean duration = 9.52 years. sessions at weekly intervals Global Panic Rating (GPR) post intervention: GPR 18.78 reduced to 4.78; PASQ
Groups: Arm 1) REBT, Follow-up numbers 74.78 reduced to 30.67; ACQ 2.37 reduced to 1.48;
arm 2) VKD (NLP arm) HADS anxiety 15.56 reduced to 7.11; HADS depression
9.67 reduced to 4.89
Pre-Post Study Design (uncontrolled)
Einspruch and Design: pre-post design. Participants: people with simple NLP Interventionist training: Mark’s Phobia questionnaire 16/17 individual participants and 27/29 group
Forman24 Setting: phobia and anxiety or multiple phobias. Mean age not reported. NLP intervention: and Beck Depression Inventory participants reported reduced phobia severity at
US outpatient clinic. Assessment: 44 years. Male 29%. Mostly group or individual intervention follow-up (both P = <0.01). Group participants
Phobia baseline and 8 weeks college educated. Recruitment: not according to assessed need. reported statistically significant improvements in
reported. Baseline and number Individual: mean 2.8 sessions depression scores (M=3.26; t = 5.18, P = <0.001), 27/29
completed not reported. Analysed per person. Duration not stated. reported reduced severity.
n = 48: group, n = 31, individual Group: weekly 2-hour sessions for
n = 17. NB: Reports only those 8 weeks plus three one-to-one
who completed both baseline sessions
and follow-up assessments
Timpany25 Design: pre-post. Setting: Participants: women with NLP interventionist training: NLP Percentage of time feeling 50% of women felt a significant reduction in symptoms
New Zealand therapist office. Baseline moderate to severe morning trainer. NLP intervention: nauseous. Number of vomiting in the week after the session. Four women went from
Morning sickness and follow-up assessment. sickness. Recruitment: press advert. combination of NLP time line episodes per day. Stress feeling nauseous 100% of the time to 20% of the time
in pregnancy Follow-up time unreported Baseline n = 12, completed n =12 therapy, well-formed outcomes/ (unreported measure) and two women from 100% to 40%. 5/8 women who
analysed n =12. Single 2-hour goal setting and hypnotherapy had been vomiting noted improvement
session
Konefal and Design: pre-post. Setting: Participants: people with NLP interventionist training: not Liebowitz Social Phobia Scale Social anxiety fear T0 M = 20.3 (SE 1.8); T1 M =12.9
Duncan26 residential training course. self-reported social anxiety. 15 male stated NLP intervention: 15 skills (SE2.0); not reported, analysed n = 23 T2 M = 12.4
Denmark Assessment: baseline (T0), and 13 female), aged 20–60 years. and techniques detailed. (SE1.7). Fear avoidance T0 = 20.1 (SE1.7), T1 = 14.5
Social anxiety post intervention (T1) and Recruitment not reported. Residential 21-day programme (SE2.2), T2 = 14.0 (SE2.2). These findings were
6 months (T2) Baseline 28, completed = statistically different from T0 to T1 (P<0.001), but not
not reported, analysed n = 23 statistically significant from T1 to T2
Gray27 Design: pre-post. Participants: substances mis-users. NLP interventionist training: not Urinalysis for illegal Non-significant difference between completers and
US Setting: community. Recruitment: compulsory stated. NLP intervention: substances non-completers. Abstinence after programme:
Substance misuse Assessments: baseline attendance. through criminal visualisation, anchoring, completers = 55%; non-completers: 16%
and 16 weeks justice system. Baseline n = 127 well-formed outcomes 2-hour
of which 99 described weekly group session and two
as valid cases. Completed n = 80, one-on-one sessions over
analysed n = 99 16 weeks
Bigley et al 28 Design: pre-post. Participants: patients who NLP interventionist training: MRI Successfully completed 38 patients (76%) successfully completed MRI. A
UK Setting: NHS radiography had previously failed to undergo radiographer with NLP practitioner MRI. Anxiety measured further nine (18%) went into the scanner but image
Claustrophobic department. MRI. 24 males/26 females. training. NLP intervention: ‘Clare’s by adapted Spielberger’s was of insufficient quality. Anxiety scores significantly
patients Assessment: on day of Median age 52 years (range 17–75) fast phobia cure’: collapsing anchor, State-Trait Anxiety reduced after NLP in all participants, but no statistical
undergoing MRI NLP session and follow up Recruitment: NHS radiography stacked anchor. Single session of Inventory difference between those completing MRI or not
on day of MRI prior to scan. department. Baseline n = 50, 1 hour duration (P = 0.172). Cost saving of £319 per MRI examination
Time lag unreported completed n = 50, analysed n = 50 of MRI with NLP vs MRI under general anaesthetic
NLP = neurolinguistic programming. OR = odds ratio. RCT = randomised controlled trial. REBT = rational emotive behaviour therapy. VKD = visual kinaesthetic dissociation.

British Journal of General Practice, November 2012 e760

 Characteristics of included studies:
Five studies were randomised controlled
trials (RCTs)19–23 and five were observational/
Citations retrieved
from searches (n = 1459)
pre-post test studies.24–28 These were
conducted in the US,3 Denmark,2 UK,2
Excluded as not
relevant (n = 1345)
Croatia, Brazil, and New Zealand.
Targeted health conditions include various
Potentially relevant
abstracts (n = 114)
anxiety disorders,6 weight maintenance,
morning sickness, substance misuse,
Foreign language and claustrophobia during MRI. The five
dissertations not
obtainable (n = 21) RCTs targeted anxiety disorders,3 maternal
anxiety/child development and weight
Citations screened
(n = 93)
maintenance. Two used NLP interventions
Total excluded (n = 83)
Descriptive (n = 31) versus a no-intervention control arm
Educational (n = 14) and three compared NLP to an active
Healthy populations (n = 18) intervention. Within the five RCTs, follow-up
Not-NLP (n = 13)
Reviews (n = 7)
occurred immediately following a single
Full text papers
reviewed (n = 41) session treatment, at 1 month, 5 months,
9 months and 3 years.
Within the 10 included studies
participants were recruited broadly from
Studies included in
the review (n = 10) childcare, criminal justice, and public
and private healthcare facilities, higher
education and the press. Demographic data
about participants was poorly reported with
eight studies reporting some data on sex.
Two of these studies specifically recruited
Figure 1. Flowchart of identified studies. based on the reported study methods women only and the six remaining reported
according to the following criteria; low risk data indicating that overall 64% of study
of bias = all domains adequately met, high participants were female. Six studies
risk of bias  =  at least one domain not reported participant age, albeit inexactly,
met and uncertain risk of bias = inadequate with a range from 17–75 years and a mid-
reporting of methods.17 Risk of bias for the range estimate of approximately 40  years.
Funding pre-post studies was assessed using the One of the RCTs21 presented broader
NHS Coventry commissioned the research Downs & Black quality index score. This is a socioeconomic data and found participants
and had representation on the steering validated checklist for assessing the quality to be college educated or higher (100%),
group. The researchers had independence of randomised and non-randomised studies married (24%), and employed (56%).
from the funders in the design and in five subscales: reporting, external validity,
execution of the study. The study was internal validity (bias and confounding), and Interventions
sponsored by University of Warwick who power.18 Delivery of NLP interventions ranged from
had a research governance role in relation three studies using a single 1–2 hour
to the study. RESULTS session19,25,28 to the remainder offering 4–20
Available evidence 1–2 hour sessions. Duration of intervention
Ethical approval
A total of 1459 citations were retrieved was reported by six studies ranging
None required. using the search strategy. Of these, 93 from 4 weeks,23 21 days residential,26
Provenance titles were potentially relevant (Figure 1). 4–5 months22,27 and 12 months.20 One study
Freely submitted; externally peer reviewed. Abstracts were obtained and screened and offered group NLP.24
41 full text papers reviewed. Of the initial Six studies described the qualifications
Competing interests 93 citations, the majority were excluded and training of the interventionists and
The authors have declared no competing as they were descriptive in nature, were these included three clinical psychologists,19
interests. not NLP interventions or they involved only eight psychotherapists,21,23 a certified NLP
Acknowledgements healthy populations. In total, 10 studies practitioner22,25 and an NLP practitioner level
We are grateful to the many NLP were identified meeting the inclusion radiographer.28 These NLP interventionists
practitioners who, in person and online, criteria (Table 1). Due to the small number were all certified to a minimum of NLP
helped us develop an understanding of of studies identified and heterogeneity practitioner level with two stated as
NLP and directed us to possible sources (in research design, populations, NLP being master practitioner certified.21,25
of evidence. interventions, and assessed outcomes), The interventionist training level was
statistical analysis was not appropriate and better described in the RCTs. Intervention
Discuss this article a narrative synthesis of the evidence was techniques reported were mixed with
Contribute and read comments about undertaken. Nine studies were published six studies19,23,25–28 employing a well-
this article on the Discussion Forum: in peer-reviewed journals and one was established NLP intervention described in
http://www.rcgp.org.uk/bjgp-discuss identified online. the NLP literature.29,30 Four studies referred

e761 British Journal of General Practice, November 2012

 (Table 2). In two RCTs the risk of bias
Table 2. Assessment of risk of bias was uncertain.19,21 None reported results
by intention to treat (ITT) analysis and,
Adequate Blinding Incomplete Overall
although one22 reported undertaking ITT
sequence Allocation (of outcome outcome data assessment
RCTs generation? concealment? assessors)? addressed? of risk of bias
analysis, only the completer analysis was
presented. Three RCTs fared better in
Krugman et al19 Unclear Unclear Yes Unclear Uncertain
reporting withdrawals and participants lost
De Miranda et al20 Unclear Unclear Unclear No High to follow-up. In the pre-post study designs,
Stipancic et al21 No No Unclear Unclear High the quality index scores18 ranged from 6–13
Sorensen et al 22
Unclear Unclear Unclear Yes Uncertain (maximum rating is 23 for non RCTs) where
Simpson and No No Unclear Unclear High low ratings represent poor quality. Only one
Dryden23 paper28 scored above the scale mid-point
for quality.
Internal
External validity: bias
Pre-post Reporting validity & confounding Power Total score NLP effects
study designs (10 items) (3 items) (13 items) (1 item) (27 items) Across the five RCTs, NLP was evaluated
Einspruch and 3 1 2 0 6
with undergraduate students, mother and
Forman24 child pairs, weight challenged adults, and
emotionally-distressed adults for which
Timpany25 2 1 2 0 5
the main outcomes were assessed ranging
Konefal and 6 0 4 0 10 from immediately post-treatment to 3-year
Duncan26
follow-up. Main outcomes reported were
Gray27 2 2 4 0 8 anxiety (self-report), child development
Bigley et al28 8 1 4 0 13 (observed), weight (objective), and quality
RCT = randomised controlled trial. of life (self-report). Four RCTs reported
no significant between group differences
in the assessed outcomes with the fifth21
generally to ‘NLP techniques’ or ‘NLP reporting less psychological distress and
behaviour modification’. The observational increased perceived quality of life in the NLP
studies provided greater detail about the group compared to the waiting list control
interventions employed. arm. Three RCTs and five pre-post studies
reported within group improvements. Of
Outcomes the three studies measuring objective
Across the 10 studies, 23 measures outcomes, one reported a post-treatment
were used, and 18 outcomes reported. increase in completed MRI scans28 and
Outcomes were largely aligned to the the other two reported no post-treatment
targeted condition and the most common improvement in urinalysis for illegal
outcome assessed was anxiety19,23–26,28 with substances27 or weight maintenance.22
three also measuring quality of life23 and
depression.23,24 Validated measures were DISCUSSION
referenced by seven studies with a further Summary
two reporting the outcome was assessed This systematic review demonstrates that
but not how. All psychological outcome there is little evidence that NLP interventions
measures were different across these improve health-related outcomes. The study
eight studies. Only three studies measured conclusion reflects the limited quantity and
objective outcomes, weight,22 successful quality of NLP research, rather than robust
completion of MRI scan,28 and urinalysis evidence of no effect.
for illegal substances.27 Of the 18 outcomes
reported, 11 were self-reported, three were Strengths and limitations
objective measures, two were observed, This represents the first well-conducted
and two not reported. The two observed review investigating the effectiveness of
measures were from RCTs19,20 and one NLP on health-related outcomes. The
had a blinded outcome assessor19 and the study has not attempted to define NLP
other was unclear.20 In general the RCTs and its many components and techniques
performed no better than the pre-post and this complicates the interpretation of
studies in terms of reporting of outcomes the evidence. This study took the authors’
and the process of their measurement. word that they were delivering NLP if they
said they were and the evidence of levels of
Risk of bias training of the interventionist supported this
In three of the RCTs, the risk of bias was assumption. Some academic investigation
high with alternate group allocation,21,23 into NLP was found in unpublished German
and incomplete outcome data reporting20 language dissertations that the library

British Journal of General Practice, November 2012 e762

 advised would not be possible to obtain methods. It was not possible to determine
and this represents a possible gap in the the risk of bias associated with selective
evidence. The decision was taken to exclude outcome reporting due to the absence of
studies using single NLP techniques. NLP published study protocols. Assessment of
has a lack of consensus surrounding a the pre-post studies found four scoring
definition of techniques and mechanism of lower than the scale midpoint score
effect and on an individual technique basis indicating a high risk of bias.
there is overlap with more established and
evidence-based psychological techniques. Implications for research and practice
Arguably these could include developing There is currently insufficient evidence to
rapport  =  person-centred counselling; recommend use of NLP for any individual
modelling  =  vicarious learning; eliciting health outcome. Neither this review, nor
well formed outcomes  =  goal setting; the FOI NHS trust data, point strongly
reframing  =  cognitive behavioural therapy to appropriate populations for further
(CBT) techniques; and anchoring = classical research. Use of NLP in specific settings
conditioning. Inclusion of studies labelled may be vindicated in future, and preliminary
by their authors as NLP and focusing on data from its use in MRI/claustrophobia may
one of these single NLP techniques would justify a sufficiently powered RCT to clarify
have lead to a misleading observation of its role for these patients. Discussions with
the evidence. Publication bias assessment NLP key informants identified populations,
was not formally calculated because only 10 for example allergy sufferers, who they felt
studies were included.18 The scoping reviews were a strong target population for further
around the practice of NLP in physical NLP-based research. A formal stakeholder
and mental health conditions suggested it consultation with a range of NLP master
remains a controversial intervention. As only practitioners would be an important next
one of the five RCTs showing a positive effect step for identifying such target populations
in favour of NLP was found, the authors for research. The strength of evidence
are less concerned about publication bias. for CBT would suggest it as a possible
However, it is possible that the controversy comparison group. The risk of bias
surrounding NLP may lead to a publication assessments point to the need to develop
bias against studies that find a positive effect a fully-specified and replicable intervention
in favour of NLP. protocol for evaluation in a sufficiently
Risk of bias in the five RCTs was high, or powered RCT.
uncertain due to inadequate reporting of

e763 British Journal of General Practice, November 2012

REFERENCES Helm, 1988: 268–280.
16. Devilly GJ. Power therapies and possible threats to the science of psychology
1. Grindler J, Bandler R. The structure of magic: a book about language and and psychiatry. Aust N Z J Psychiatry 2005; 39(6): 437–445.
therapy. Oxford: Science and Behaviour, 1979. 17. Higgins J, Green S (eds).The Cochrane handbook for systematic reviews of
2. Bandler R, Grinder J. Frogs into princes. Moab, UT: Real People Press, 1979. interventions. [updated March 2011]. The Cochrane Collaboration, 2011. http://
www.cochrane-handbook.org/ (accessed 3 Aug 2012).
3. The Association for NLP. Accreditation panel. http://www.anlp.org/
accreditation-panel (accessed 8 Oct 2012). 18. Downs SH, Black N. The feasibility of creating a checklist for the assessment
of the methodological quality both of randomised and non-randomised
4. International NLP Trainers Association. Standards. http://www.inlpta.org/ studies of health care interventions. J Epidemiol Community Health 1998;
index.php/en/standards-mainmenu-77 (accessed 8 Oct 2012). 52(6): 377–384.
5. Camana R. Depression: can NLP relieve the sypmptoms of depression? 19. Krugman M, Kirsch I, Wickless C, et al. Neuro-linguistic programming
Articlesantch.com http://www.articlesnatch.com/Article/Depression---Can- treatment for anxiety: magic or myth? J Consult Clin Psychol 1985; 53(4):
Nlp-Relieve-The-Sypmptoms-Of-Depression-/1808517#ixzz22U7gGqJj 526–530.
(accessed 3 Aug 2012).
20. de Miranda C, de Paula C, Palma D, et al. Impact of the application of
6. Russell M. ‘Healer of last resort’ — if you’re caught in the health practitioner neurolinguistic programming to mothers of children enrolled in a day care
round-about. NSW, Australia: Natural Therapy Pages. http://www. centre of a shantytown. Sao Paulo Med J 1999; 117(2): 63–71.
naturaltherapypages.com.au/therapist/weightlossadelaide/8722 (accessed 3
21. Stipancic M, Renner W, Schutz P, Dond R. Effects of neuro-linguistic
Aug 2012).
psychotherapy on psychological difficulties and perceived quality of life.
7. Thomson G (ed.). Magic in practice: introducing medical NLP — the art and Counselling and Psychotherapy Research 2010; 10(1): 39–49.
science of language in healing and health. London. Hammersmith Press, 2008. 22. Sorensen LB, Greve T, Kreutzer M, et al. Weight maintenance through
8. Acuity Training and Development. Advanced NLP for doctors. http://www. behaviour modification with a cooking course or neurolinguistic
acuitydr.co.uk/certified-nlp-core-skills-training (accessed 3 Aug 2012). programming. Can J Diet Pract Res 2011; 72(4): 181–185.
9. Jo Wadell Training. NLP foundation course for doctors. http://www.jowaddell. 23. Simpson SDR, Dryden W. Comparison between REBT and visual/kinaesthetic
co.uk/2012-NLP-doctors-course-details.php (accessed 3 Aug 2012). dissociation in the treatment of panic disorder: an empirical study. J Rat-Emo
10. Medical NLP. http://medicalnlp.groupsite.com (accessed 8 Oct 12) Cognitive-Behav Ther 2011; 29(3): 158–176.
11. Ismail K, Winkley K, Rabe-Hesketh S. Systematic review and meta-analysis 24. Einspruch E, Forman B. Neurolinguistic programming in the treatment of
of randomised controlled trials of psychological interventions to improve phobias. Psychother Priv Pract 1988; 6(1): 91–100.
glycaemic control in patients with Type 2 diabetes. Lancet 2004; 363(9421): 25. Timpany L. A study of the effectiveness of single session NLP treatment for
1589–1597. pregnancy sickness. 1994. http:/www.transformations.net.nz (accessed 3 Aug
2012).
12. Alam R, Sturt J, Lall R, Winkley K. An updated meta-analysis to assess
the effectiveness of psychological interventions delivered by psychological 26. Konefal J, Duncan R. Social anxiety and training in neurolinguistic
specialists and generalist clinicians on glycaemic control and on programming. Psychol Rep 1998; 83(Pt 3): 1115–1122.
psychological status. Patient Educ Couns 2009; 75(1): 25–36. 27. Gray R. The Brooklyn programme — innovative approaches to substance
13. Lamb SE, Hansen Z, Lall R, et al. on behalf of the Back Skills Training Trial abuse treatment. Fed Prob: a Journal of Correctional Philosophy and Practice
investigators. Group cognitive behavioural treatment for low-back pain in 2002; 9–16.
primary care: a randomised controlled trial and cost-effectiveness analysis. 28. Bigley J, Griffiths PD, Prydderch A, et al. Neurolinguistic programming used
Lancet 2010; 375(9718): 916–923. to reduce the need for anaesthesia in claustrophobic patients undergoing
14. National Institute of Health and Clinical Excellence. Depression: the treatment MRI. Br J Radiol 2010; 83(986): 113–117.
and management of depression in adults. Clinical guideline 90. London: NICE, 29. Dilts R, Delozier J. The encyclopaedia of systemic NLP and NLP new code
2009. literature. NLP University Press. http://www.nlpu.com (accessed 3 Aug 2012).
15. Heap M. Neurolinguistic programming: an interim verdict. In: Heap M (ed.). 30. James T, Woodsmall W. Time line therapy and the basis of personality.
Hypnosis: current clinical, experimental and forensic practices. London: Croom Capitola, CA: Meta Publications, 1988.

British Journal of General Practice, November 2012 e764