METHODOLOGICAL ISSUES IN MEASURING COERCION by 6DJgJytp

VIEWS: 6 PAGES: 36

									                             HOW




           Georg Høyer
Institute of community medicine
      University of Tromsø
        Holmen, March 5th 2011
→ We still lack a valid measure for coercion

→ We still don’t know exactly what coercion is, how it shall be
 defined and eventually how to grade and measure coercion

→ Which again is why we lack knowledge about the
 effect of coercive interventions

→ There are reasons to believe that PERCEIVED
COERCION is the most important element in this
perspective, but also the most difficult to measure

                       Holmen, March 5th 2011
“...What makes people feel coerced - is a prerequisite to
understanding coercion as an independent variable (i.e., whether
and how coerced hospitalization is effective in producing therapeutic
outcomes)”                             Monahan et al. 1995




                           Holmen, March 5th 2011
               WHAT CONSTITUTES
                 COERCION?
No-choice coercion         FORMAL              VIOLATION OF INTEGRITY
                                               AND/OR AUTONOMY
(rotten choices)           (LEGAL)
                          COERCION             HUMILIATION
Structural coercion
                                               OTHER FACTORS
(ward rules etc)



                      PROCEDURES



   PHYSICAL
  (CONCRETE)                                           PERCEIVED
   COERCION                                            COERCION
                      Holmen, March 5th 2011
HOW CAN PERCEIVED COERCION BE EXPLORED?
           how
Instruments/questionnaires
      Self-administered (Postal, on-site)
      Filled in by interviewer (Telephone, video, on-site)

       Global assessments vs structered, multidemential instruments

Qualitative interviews

Clinical interviews

                         Holmen, March 5th 2011
MILESTONES IN THE EXPLORATION OF
PERCEIVED COERCION
           how

Late 1970th : First publications on patients’ experiences

1992: The MacArthur coercion study. Development of the AES
      and the MPCS

1997: Introduction of the Coercion Ladder (CL)




                           Holmen, March 5th 2011
                  VARIOUS INSTRUMENTS

1978-1995: Mostly self-designed questionnaires, rarely used in
                   how
           more than one study

1995: The MacArthur Perceived Coercion Scale (MPCS), (Lidz et al, 1995)

1997: The Coercion Ladder (CL), (Høyer et al., 2002)

2001: The Community Perceived Coercion Scale, (Birmbaum, Lidz & Greenidge 2001)

2005: Psychiatric Experience Questionnaire (PEC), Frueh et al, 2005

2006: Perceived Coercion in Everyday Life (PCEL), (Steadman & Redlich, 2006)

2010: Coercion Experience Scale (CES), (Bergk, Flammer & Steinert, 2010)

                                Holmen, March 5th 2011
                      AES/MPCS


              how
MPCS developed from a 104 item semistructured interview
schedule, The Admission Experience Interview (AEI), through
a 41 item questionnaire, The Admission Experience Survey, to a
15 (or 16) item version, The Admission Experience Scale, The
AES

The AES consists of 3-4 subscales (often given different names),
one of them being the MPCS. Others are Voice (or “process
exclusion”) (4 items), Negative pressures (or “Force/Threats”)
(6 Items).

                       Holmen, March 5th 2011
               THE MACARTHUR PERCEIVED
                COERCION SCALE (MPCS-5)


• (1) I had more influence than anyone else on whether
  I came into the hospital (Influence)
• (4) I had a lot of control of whether I went into the
  Hospital (Control)
• (7) I chose to come into the hospital (Choice)
• (14) I felt free to do what I wanted about coming into
  the hospital (Freedom)
• (15) It was my idea to come into the hospital (Idea)

                     Holmen, March 5th 2011
             MPCS-5: Some concerns

• The definition of coercion: Coercion defined as
  lack of (or reduced) autonomy

• The terms influence, control, choice and freedom (and
  idea) were chosen to constitute perceived coercion
  because it proved difficult to ask someone directly
  about coercion (The terms were chosen on basis of
  their face validity as everyday synonyms for
  autonomy)

• If patients tells us that influence, control and the like,
  were absent, then coercion was present (Gardner et al 1993)

                       Holmen, March 5th 2011
              MPCS-5: Some concerns II

• Validation problems (No ”Gold Standard”)
• Studies on the reliability of perceived coercion measures almost
  non-existent
• Low impact of the application of coercive measures on
  perceived coercion
• Focus on the admission situation only (and does not
  discriminate between what happens in the community and at
  arrival to the hospital)
• Cultural, socioecconomic, gender and race sensitive
• Different scoring procedures (little discussed) (True/False,
  yes/no, Lickert score)
                         Holmen, March 5th 2011
               MPCS-5: Some concerns III
Not very user friendly?
              % completion of AES/MPCS/CL
              AES        MPCS          CL
              65.4       72.8          95.8
                                                    (The Nordic coercion study)


Missing rate (%) for individual questions in MPCS: 1.6-11.8

Gardner et al. 1993: MPCS-5:11.8 %, but MPCS-4: 2.5-3.7

Nicholson et al. 1997: More than 20% refused to fill in the AES
                           Holmen, March 5th 2011
                 MPCS-5: Some concerns III
Low test-retest reliability (consistancy)
Number of inconsistent answers when the four ”voice” questions were
repeated during the same interview (%) (The Nordic coercion study, the
Danish subsample)



1 (AES 3):      14.8 %               COMBINED:
2 (AES 5):      14.8%
                                     INCONSISTENT: 51.9%

3 (AES 9):      22.2%                 INCONSISTENT 68.5%
                                      + MISSING
4 (AES 13):      12.9%
                             Holmen, March 5th 2011
         Perceved coercion: Other concerns

Relations between Perceived Coercion and:
     Patient Satisfaction
     Humilation
     Violation of integrity
     Trauma
     Quality of Life
     Symmptom measures

                  Holmen, March 5th 2011
                          WHAT DO WE KNOW ABOUT
                           PERCEIVED COERCION?

There is a tendency that patients either feel coerced or not, and not a
straight ”dose-effect” response in perceived coercion




                           Holmen, March 5th 2011
                        THE NORDIC COERCION STUDY
                        DISTRIBUTION OF MPCS SCORES

                   Legally involuntary                            Legally voluntary
           140
                                                           120


           120
                                                           100

           100

                                                            80
           80

                                                            60
           60


           40                                               40



           20
                                                            20
Count




                                                   Count



            0
                    0    1   2   3   4     5                 0
                                                                    0   1   2   3     4   5
                 MPCS
                                                                 MPCS


        MPCS-5 Scores                    Holmen, March 5th 2011
The Nordic Coercion Study
Mean scores on the MPCS-5
                          MPCS
                        ALL Vol Invol

The Nordic Study         2.5          1.7 3.5
Bindman                  2.6          1.9 3.4
Hoge (1978)               -           0.6 3.2
Hiday (1997)             2.9           -   -




             Holmen, March 5th 2011
                     WHAT MORE DO WE KNOW ABOUT
                         PERCEIVED COERCION?


Gender, Age, Diagnosis, Degree of symptoms, Formal legal status,
Number of previous admissions, are rarely associated with
perceived coercion (with a few exceptions)

Procedural justice/negative pressures/process exclusion/voice are
the most important predictors of perceived coercion

Surpirsingly low correspondence between use of physical coercion
and perceived coercion



                          Holmen, March 5th 2011
                      WHAT MORE DO WE KNOW ABOUT
                         PERCEIVED COERCION II



Perceived coercion scores seem to be stable over time, even if more
patients agree that the commitment was necessary as times go by

More than half of the committed patients feel they have recieved
help and have been treated well by the staff




                          Holmen, March 5th 2011
       Of course no-one, but our
       Honourable guest speaker

       Professor Chuck Lidz

       University of Massachutes
       Medical School, USA




Holmen, March 5th 2011
       PERCEIVED COERCION:
    SOME EMPIRICAL RESULTS VII
What happened in the community before hospitalization was the best predictor
of perceived coercion measured (> 2 days after admission)
                                                  Cascardi & Poytress 1997

56.4 % af all committed patients said they would have accepted an offer to be
admitted voluntarily
                                                          Hoge et al, 1997

20-30 % of patients receiving ECT reported that they did not have the
opportunity to say no even if this procedure required informed consent
                                                           Rose et al, 2005

44% of voluntarily admitted patients beleived they would be formally detained if
they tried to leave the ward
                              Holmen, March 5th 2011      Bindman et al, 2005
PERCEIVED COERCION: SOME EMPIRICAL
RESULTS VI: Restraint and perceived coercion
18 % of the committed patients had been subjected to physical
coercion. However, the use of physical coercion was NOT significantly
correlated to perceived coercion
                                                Iversen et al 2007

10 % subjected to physical force, low correlation to perceived coercion (0.27)
                                                           Lidz et al, 1998

19/138 were subjected to restraint and 29/138 were secluded. Restraint was a
significant predictor of high perceived coercion scores (p<0.02)
                                                           McKenna et al, 1999

In other words:
Surpirsingly low correspondence between use of physical coercion an
Perceived coercion
                              Holmen, March 5th 2011
    PERCEIVED COERCION:
  SOME EMPIRICAL RESULTS II

50 % of involuntarily and 40 % of voluntarily admitted
patients said their inetgrity had been violated
                                                Kjellin et al 1996




                       Holmen, March 5th 2011
PERCEIVED COERCION: SOME EMPIRICAL RESULTS VI             : Accounts
of being subjected to forced medication or restraint (%)


                 Registered Patients’ Relatives Head nurses’
                            reports reports     report
Legally invol.
patients            23             65             45           22
Legally vol
patients             0             28             10            3

                                                  Kjellin & Westrin 1998
                         Holmen, March 5th 2011
                     VALIDITET
Innholdsvaliditet (content validity): Instrumentet (variabelen) må
inneholde (alle) viktige faktorer av betydning for det man ønsker å
måle, og variablene må ha et hensiktsmessig format

Kriterievaliditet (Criterion validity): Korrelasjon mellom
instrument (variabel) og gullstandard. Concurrent validity og
predictive validity er begge varianter av kriterievaliditet

Konstruksjonsvaliditet (construct validity): I hvilken grad
sammenfaller resultatene instrumentet gir med empiri som ligger til
grunn for den teoretiske konstruksjon av begrepet (fenomenet) vi vil
undersøke. Må brukes når vi ikke har noen gullstandard
                               Sitzia J. J for quality in health care 1999; 11(4):319-24



                           Holmen, March 5th 2011
Konstruksjonsvaliditet (construct validity) fortsatt:

Tre fremgangsmåter:
1. Korrelasjon mellom instrumentet og andre teoretiske og
    observerbare (målbare) mål for samme fenomen

2. Analyse av empiriske data for å se hvilke andre fenomen det
   aktuelle instrumentet korrelerer med

3. Anvende instrumentet for å se om det fungerer som forventet.
      For eksempel skiller instrumentet klart mellom to grupper som
      man ville forvente befant seg i hver enda av skalaen, dvs
      instrumentets evne til å diskriminere
eller
      Se på graden av korrelasjon mellom instrumentet og de
      variablene det burde og burde ikke korrelere med
                             J. J for March 5th 2011
                      Sitzia Holmen, quality in health care 1999; 11(4):319-24
                    VALIDITET

Intern validitet: Relaterer seg til tilfeldige feil, systematiske
feil og confounding

Ekstern validitet: Kan resultatene generaliseres og har de
praktisk betydning?

                                  Bjørndal og Hofoss 2004




                         Holmen, March 5th 2011
                  AES-15: MPCS
Number of questions answered:

                                                 1. (AES 1): 15 missing
  All five: 563
                                                 2. (AES 4): 11missing
  Four questions: 649                            3. (AES 7): 9 missing
                                                 4. (AES 14): 24 missing
                                                 5. (AES 15): 27 missing
  Three or less: 701




                        Holmen, March 5th 2011
PERCEIVED COERCION; PREDICTORS
  Model 1: AES-factors NOT included
MPCS                                     COERCION LADDER
Legal status                             Legal status
                                         Own opinion of legal status
Own opinion of legal status
                                         Offended during the admission
Own idea to be admitted
                                         Good to be admitted
p<0.05       R2= 0.26
                                         GAF
                                         p<0.05         R2= 0.22



                          Holmen, March 5th 2011
         Perceived Coercion; PREDICTORS
         Model 2: AES-factors included
         (Linear regression, backwards)
MPCS                                      COERCION LADDER

Negative pressures                        Negative pressures
Process exclusion                         Process exclusion
Own opinion of legal status               Own opinion of legal status
Own idea to be admitted                   BPRS-16
Necessary to be admitted
                                           p<0.05        R2=0.61
p<0.05       R2= 0.61
                           Holmen, March 5th 2011
       The Nordic Coercion Study
           logistic regression
Low-High perceived coercion Low-High perceived coercion
(MPCS 0-3 vs 4-5)           (Ladder 1-4 vs 5-10)
                 OR          CI                               OR         CI
Own idea                                    Own idea
to be admitted   14.38   5.71-36.22         to be admitted    3.90   1.72-8.87

                                            Offended during   0.34   0.17-0.67
Negative                                    admission
Pressures         1.44   1.10-1.91
                                            Negative
Process                                     Pressures         1.24   1.24-1.89
exclusion         2.30   1.66-3.19
                                             Process
                              Holmen,   Marchexclusion
                                             5th 2011         1.27   1.27-2.08
AES-15




Holmen, March 5th 2011
                 Legal status: voluntary
                      Denmark Finland Norway Sweden Iceland
The patients’
reports: Came to         n=47   n=107   n=96 n=133 n=114
the hospital …             %       %      %      %       %
Involuntarily                12.8           7.5          16.7    5.3    9.6
Voluntarily                  74.5         87.9           81.3   85.7   86.0
Neither/don’t
know/no answer               12.8           4.7           2.1    9.0    4.6
Chi-Square = 16.91, df = 6, p = 0.010



                                Holmen, March 5th 2011
              Legal status: involuntary
                      Denmark Finland Norway Sweden                    Iceland
The patients’
reports: Came to         n=48   n=122  n=162   n=93                        n=8
the hospital …             %       %      %      %                           %
Involuntarily               77.1           58.2          50.6   66.7      100
Voluntarily                   6.3          32.8          46.9   19.4        0
Neither/don’t
know/no answer              16.7             9.0          2.5   14.0        0

   Chi-Square = 45.25, df = 6, p = 0.000


                                Holmen, March 5th 2011
          The Nordic Coercion Study
            Study Sample Level 3
           Interviewed   Completed      Completed   Completed   Completed
                           AES          MPCS(5)        CL       MPCS&CL
Denmark       91            35              50         83          35
Finland       224            -               -        204           -
Iceland       131           94              99        128          93
Norway        253          214             228        250         213
Sweden        229          163             186        215         159
Total         928          506             563        880         500
                           Holmen, March 5th 2011
Holmen, March 5th 2011

								
To top