C International Psychogeriatric Association 2015 International Psychogeriatrics (2016), 28:3, 373–384 doi:10.1017/S1041610215001490
REVIEW
Decision-making capacity evaluation in adult guardianship: a systematic review ...........................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................
Sally McSwiggan, Susanne Meares and Melanie Porter Department of Psychology, Macquarie University, North Ryde, Australia
ABSTRACT
Background: Evidence of impaired decision-making capacity is a legal requirement for adult guardianship. To understand the quality of the evidence health professionals commonly provide in reports submitted to guardianship courts, a systematic review was undertaken to appraise the design and methodological quality of the published literature on health professionals’ written reports of decision-making capacity and to describe the content of these reports. Methods: Electronic searches from 1980 to 2015 identified 1183 articles of which 11 met the inclusion criteria where each evaluated quantitatively the content of health professionals’ written reports submitted to adult guardianship proceedings. Methodological quality of the selected studies was rated using a critical appraisal tool. Results: Nine studies sourced files from courts in the U.S. and one from Sweden; another reported on guardianship decisions from Australia. Four studies were rated as moderately strong or strong. Strengths included the use of comparison groups and a reliable and valid instrument to code reports. The review showed a person’s medical condition was often cited as evidence of impaired decision-making capacity. Cognitive, psychiatric and functional abilities were less often described, and a person’s values and preferences were rarely recorded. Conclusions: It is recommended health professionals describe the process by which a person makes a particular decision (their ability to understand, appreciate, reason and communicate) in addition to providing medical information, including cognitive, psychiatric and functional abilities. This approach provides support for a professional’s opinion and evidence for a court. International studies of health professionals’ approach to decision-making capacity evaluation are needed. Key words: decision-making capacity, competency, adult guardianship, systematic review
Introduction Capacity refers to a judgement as to whether an individual has the requisite cognitive, decisional, affective, and practical skills to be found to have the ability to adequately complete a particular task (such as driving), or to make a particular decision (such as to refuse medical treatment) (Moye and Braun, 2010). Decisional abilities are distinct from practical (also known as executional) abilities in so far as the person may not be able to accomplish Correspondence should be addressed to: Sally McSwiggan, Institute for Biomedical Ethics, Basel University, Bernoullistrasse 28, Basel 4056, Switzerland. Phone: +41 (0) 61 267 17 86, +41 (0) 79 244 75 78; Fax +41 (0) 61 267 17 80. Email: [email protected]. Received 24 Mar 2015; revision requested 13 Jul 2015; revised version received 17 Aug 2015; accepted 18 Aug 2015. First published online 28 September 2015.
their choice without assistance (Collopy, 1995; Boyle, 2008). Although conclusions about capacity are ultimately determined by the law, in practice most views formed about a person’s capacity are established by health professionals, lawyers, social care workers, friends and family (Kapp, 2002; Bennett and Hallen, 2005). When a view is formed that a person has impaired capacity, the vast majority of solutions are informal such as family or friends offering assistance. However, when informal solutions cannot be found or issues of conflict are raised, the law may have a role. Applications are made to guardianship (also known as conservatorship) courts and tribunals seeking legal orders to appoint a substitute decisionmaker, also known as a guardian or conservator (Leatherman and Goethe, 2009). A critical piece
374
S. McSwiggan et al.
of evidence in the legal proceedings is the clinical report of decision-making capacity written by a health professional (Moye et al., 2005). Models of decision-making capacity that have developed over the last 30 years have been based on empirical research, expert consensus and case law (Appelbaum and Roth, 1981; Grisso and Appelbaum, 1998; Moye and Marson, 2007; Karlawish, 2008). Key components include: (1) the presumption of capacity, (2) capacity as domain specific to the lifestyle area under review (e.g., finances or personal care), (3) decision specific within that domain (e.g., paying bills or selling real estate), and (4) temporally limited due to the transient and progressive nature of some disabilities (e.g., delirium or Alzheimer’s disease) (Ganzini et al., 2003; Moberg and Kniele, 2006; Boyle, 2008; Newberry and Pachet, 2008). Research has helped to advance the understanding that decision-making capacity is not a global, “all or nothing” status but rather a situation-specific functional definition (Anderer, 1990; Sabatino and Basinger, 2000; Peisah et al., 2009). Reviews of empirical research grouped by cognitive and psychiatric disorders have reliably found that medical diagnoses alone do not predict impaired decision-making capacity (Okai et al., 2007; Moye et al., 2006; van Staden, 2009). More specifically, no single neuropsychological test, cognitive domain or psychiatric presentation consistently predicts a lack of capacity (Okai et al., 2007; Palmer and Savla, 2007; Moberg and Rick, 2008). Additionally, mental status screening has little utility beyond staging cognitive impairment (Pachet et al., 2010). The results show that decision-making capacity is more than group membership or scores achieved on standardized tests. Surveys of health professionals have suggested decision-making capacity is often considered challenging to assess and report, commonly due to limited knowledge of the concept (Ganzini et al., 2003; Hill et al., 2006; Mullaly et al., 2007). To aid health professionals in providing clinical reports of decision-making capacity for adult guardianship courts and tribunals, there has been an emergence of resources (e.g., Ontario Ministry of the Attorney General, 2005; Moye et al., 2007a; American Bar Association Commission on Law and Aging and American Psychological Association, 2008; Australian Psychological Society and the Office of the Public Advocate, 2009; British Psychological Society, 2010; O’Neill and Peisah, 2011). To date, however, no systematic review has been undertaken of the literature that has examined the content of written reports of decisionmaking capacity tendered by health professionals for adult guardianship proceedings. The aims of this systematic review were (a) to critically appraise
the design and methodological quality of the published literature on decision-making capacity reports written by health professionals for adult guardianship proceedings and (b) to describe the content of these reports.
Methods Searches were conducted using three electronic databases: Ovid MEDLINE (1948 to February, 2015), PsycINFO (1967 to February, 2015) and Scopus (1980 to February, 2015) from 1980 until February 1, 2015 (Figure 1).The search terms were developed based on specific key words to capture guardianship proceedings (guardianship/ guardian/conservatorship/conservator) and capacity (capacity/competency/decision making/decision making capacity/incompetent/incompetence/incompetency/mental capacity/mental incapacity) or disability (disability/dementia/brain injury/cognition/ cognitive impairment) or proceedings (proceedings/determination/order/finding/evidence). The selection criteria for inclusion in the systematic review were as follows: (a) the article was written in English, (b) participants were aged 18 years or above, (c) capacity reports for guardianship proceedings were evaluated and (d) the article reported results using quantitative analyses. The exclusion criteria were as follows: (a) opinion articles and narrative reviews, (b) studies containing less than 10 subjects. Articles meeting the selection criteria were rated for methodological quality in 11 key areas using a 14-item critical appraisal tool developed by Heacock et al. (1997; Table 1). Scores ranged from 0–12. Heacock and colleagues (1997) categorized scores as strong (9–12 points), moderately strong (more than 6 and up to 9 points) and weak (6 points and less). The critical appraisal tool was developed for non-experimental quantitative research. It was selected for use in the current study because a guide to scoring each item was provided, and the scale had undergone content validation and reliability testing (Crowe and Sheppard, 2011). Heacock and colleagues (1997) indicate good to excellent interrater reliability of the checklist items apart from the first item which was reported as poor. Each of the key areas was weighted equally in terms of importance for methodological quality and was assigned a value of one point in the checklist with one exception being area 9 (Appropriateness of Statistical Test/Sample Size) that was assigned a value of two because it was considered vital to the valid interpretation of research results (Heacock et al., 1997). This weighting scheme attempts to address the problem whereby summary scores hide defects if studies score well in some areas considered
Decision-making capacity in adult guardianship
375
Figure 1. Flowchart of the search strategy.
not as methodologically important as others (Crowe and Sheppard, 2011). Results of the database searches were downloaded and imported into Endnote files. Duplicate articles were excluded. The title and abstracts of all articles were reviewed by the first author to identify those meeting selection criteria. Where abstracts were not available or study methods were ambiguous, the full text of the article was sourced and examined. The full texts of all articles meeting selection criteria were retrieved. Reference lists of the articles meeting the selection criteria, opinion papers and narrative reviews were searched for additional relevant articles.
Results The results of the searches produced 961 articles from Ovid MEDLINE and PsycINFO and 222 from Scopus. Ten articles met the inclusion criteria. Another article was sourced by examining the references from the selected articles, bringing the total to eleven. Figure 1 provides details of the identification process, including the reasons for exclusion.
Study quality Research questions examining reports of capacity written for guardianship proceedings are restricted in so far as experimental designs cannot be applied due to the ethical and legal constraints imposed by the lawful process being examined. These constraints limit research designs to observational studies which are largely descriptive. Of the articles selected, ten used retrospective cross-sectional designs and one was a before and after design (Keith and Wacker, 1993). Capacity reports were generally sourced from guardianship files held by the courts. As shown in Table 2, nine studies sourced files from between one and 30 courts across various states in the U.S. and one other sourced files from a court in Sweden. Reports were variously selected by randomizing guardianship files (Stevenson and Capezuti, 1991), choosing alphabetically-consecutive files to meet a quota (Bulcroft et al., 1991), convenience samples (Jurectic et al., 1993) and every other consecutive application (Björkstén et al., 2014). The ten studies examined in total 2443 guardianship files from the U.S. and Sweden. The remaining study reported on 285 reasons for decision (also known as judgements) from guardianship hearings in Australia (see Table 2).
376
BARRITT STEVENSON KEITH
KEY AREAS
ITEMS
LISI AND
DUDLEY
JURETIC BARINAGA- AND
MOYE
BJÖRKSTÉN
PETERS BULCROFT AND
AND
ET AL., ET AL.,
CAPEZUTI,
WACKER, ET AL.,
BURCH,
GOINS,
ET AL., PARKER, GIBSON, ET AL.,
1985
1991
1993
1995
2003
2007
1991
1993
2008
2011
2014
...........................................................................................................................................................................................................................................................................................................................................................................................................................................................
Gap in literature Research question Methodology Measurement Representativeness Representativeness Control/comparison group Sample size/power Statistical test(s) Limitations/bias Limitations/bias Statistical test/sample size Implicationsa Implicationsa
1 2 3 4 5 6 7 8 9 10 11 12 13 14 Total
1 1 0 0 0.5 0.5 0 0 0 0.5 0 2 0 0 5.5
1 1 1 0 0.5 0.5 0 0 0 0 0 2 0 0 6
0 1 1 0 0.5 0.5 0 0 0 0 0 2 0 0 5
1 1 1 0 0.5 0.5 1 0 1 0 0 2 0 0 8
0 1 1 0 0.5 0.5 0 0 0 0 0 2 0 0 5
1 1 0 0 0.5 0.5 0 0 0 0 0 2 0 0 5
1 1 1 1 0.5 0.5 1 0 0.5 0 0 2 0 0 8.5
1 1 1 1 0.5 0.5 1 0 1 0.5 0.5 2 0.5 0.5 11
0 1 1 0 0.5 0.5 0 0 0 0 0 2 0 0 5
0 1 1 1 0.5 0.5 1 0 1 0 0 2 0 0 8
1 1 1 0 0.5 0.5 0 0 0 0 0 2 0 0 6
Notes: Gap in Literature: 1 = Does the study identify a gap in the existing literature; Research Question: 2 = Is the research question clearly stated; Methodology: 3 = After reading the methods section, could the reader repeat the study with confidence; Measurement Instruments/ Techniques: 4 = Do the authors state the measurement technique(s)/ instrument(s) are valid/ reliable or do they provide information so the reader can investigate the validity and reliability of the measurement technique(s)/ instrument(s); Representativeness: 5 = Is the study population representative of the end-user population; Representativeness: 6 = Is the study setting representative of the workplace to which the results will be applied; Control/ Comparison Group: 7 = Is there an appropriate control/comparison group; Sample Size: 8 = Is there a statement about sample size or power of the study; Statistical Test(s): 9 = Do the authors state the statistical test used and level of significance or confidence levels achieved; Limitations/Bias: 10 = Do the authors discuss the limitations or biases of the study design and/or methodology; Limitations/ Bias: 11 = Do the authors discuss the effects of the preceding limitations or biases on the results; Appropriateness of Statistical Test/ Sample Size: 12 = Do the authors achieve any one of the following; utilize a study sample with at least 30 subjects or; conduct non-parametric statistics tests for samples with less than 30 subjects or; conduct parametric tests for samples with less than 30 subjects but explain the sample is normally distributed or; conduct multivariate tests with ten times as many subjects as there are independent variables; Implications: 13 = Do the authors discuss recommendations of the findings; Implications:14 = Do the authors identify opportunities for further research based on the current study. Methodological ratings are from Heacock et al., (1997). a Only answer items 13 and 14 if you responded yes to item 11.
S. McSwiggan et al.
Table 1. Critical appraisal ratings of the selected articles
Table 2. Data extraction from the selected articles ARTICLE (CAT SCORE) ( C A T C L A S S I FI C A T I O N )
SAMPLE
METHOD
LOCATION
STATISTCAL ANALYSIS
RESULTS
...........................................................................................................................................................................................................................................................................................................................................................................................................................................................
N = 42 cases aged ࣙ 18 years
Bulcroft et al. (1991) (6/12) (weak)
N = 63 cases aged ࣙ 50 years
Stevenson and Capezuti (1991) (5/12) (weak)
N = 49 cases aged ࣙ 60 years
Keith and Wacker (1993) (8/12) (moderately strong)
N = 766 cases aged ࣙ 60 years exclusion: voluntary guardianship N = 40 cases aged ࣙ 60 years exclusion: acute trauma
Barritt Lisi and Barinaga-Burch (1995) (5/12) (weak)
Dudley and Goins (2003) (8.5/12) (moderately strong)
Juretic et al. (1993) (5/12) (weak)
Written reports and 8 observed hearings from one court, January, 1977 to May, 1982 Written reports from two courts, alphabeticallyconsecutive files were selected to meet a quota, January, 1987 to December, 1988 Written reports from one court; reports were selected at random, 1987
FL, U.S.
frequency
7% described symptoms of physical incapacity; 12% described symptoms of mental incapacity
WA, OH, U.S.
frequency interrater reliability
Demonstration of incompetence “rarely supplied”; most common phrase “failure to manage estate,” followed by “confusion/ disorientation”
PA, U.S.
frequency
Written reports from court files before/after legislation reform, 1984 (IA), 1983 (MO)
IA, MO, U.S.
frequency t-test
>50% medical diagnosis only evidence; 20% nonspecific diagnosis “organic brain syndrome”; 10% “senile” 14% increase in “medical conditions” described (IA) after legislative reform
Written reports from four courts selected on availability, June, 1988 to June, 1989
TX, U.S.
frequency interrater reliability
N = 726 cases aged ࣙ 60 years
Written reports and observed hearings from 30 courts, October, 1989 to January, 1992
frequency
N = 119 cases aged ࣙ 65 years exclusion: reviews of guardianship and petitions for conservatorship
Written reports from six courts, December, 1996 to December, 1999
CA, CO, FL, IN, KS, MI, MN, NY, OR, WA, U.S. PA, WV, U.S.
frequency chi-square
75% medical or psychiatric diagnosis; 8% mental status examination; 48% unsupported statements of incompetence; 8% prognosis or potential need for re-evaluation >50% medical evidence
95% (PA) versus 66% (WV) cognitive assessment; 2% overall met minimum reporting criteria in all clinical areas (cognitive, psychiatric, functional, current treatment)
Decision-making capacity in adult guardianship
Peters et al. (1985) (5.5/12) (weak)
377
378
ARTICLE (CAT SCORE) ( C A T C L A S S I FI C A T I O N )
SAMPLE
METHOD
LOCATION
STATISTCAL ANALYSIS
RESULTS
...........................................................................................................................................................................................................................................................................................................................................................................................................................................................
Moye et al. (2007b) (11/12) (strong)
N = 298 cases aged ࣙ 55 years exclusion: developmental delay
Written reports and oral transcriptions from eight courts, January, 2002 to December, 2005
MA, PA, CO, U.S.
frequency chi-square
Parker (2008) (5/12) (weak)
N = 285 reasons for decision
QLD, Australia
frequency
Gibson (2011) (8/12) (moderately strong)
N = 80 cases (240 capacity reports) aged ࣙ 50 years
Published reasons for decision, AUSTLII website, January, 2005 to April, 2008 Written reports and oral transcriptions from two courts, 2004, 2005, 2006
KY, U.S.
frequency chi-square one-way analysis of variance
Björkstén et al. (2014) (6/12) (weak)
N = 260 cases aged ࣙ 50 years exclusion: involuntary guardianship and normal memory
STK, Sweden
frequency
Written reports from one court; every other consecutive application was selected, June 2007 to July, 2008
86%, 63%, 87% medical diagnosis (MA, PA, CO); 61%, 47%, 86% cognitive abilities (MA, PA, CO); 5%, 18%, 23% mood disturbance (MA, PA, CO); 5%, 5%, 19% cognitive screening (MA, PA, CO); 1%, 2%, 34% neuropsychological assessment (MA, PA, CO) 76% of reasons contained legal findings on capacity;44% had evidence from two or more experts; 29% disagreed 88%, 95%, 94% medical diagnosis; 53%, 91%, 89% cognitive abilities; 23%, 71%, 77% functional abilities (physician, psychologist, social worker) 20% mental status testing or neuropsychological assessment
Notes: CAT = critical appraisal tool; CA = California; CO = Colorado; IA = Iowa; IN = Indiana; FL = Florida; KS = Kansas; KY = Kentucky; MA = Massachusetts; MI = Michigan; MN = Minnesota; MO = Missouri; NY = New York; OH = Ohio; OR = Oregon; QLD = Queensland; STK = Stockholm; TX = Texas; WA = Washington; WV = West Virginia.
S. McSwiggan et al.
Table 2. Continued.
Decision-making capacity in adult guardianship
The methodological rating scores for the eleven articles ranged from 5 to 11 according to the Heacock et al. (1997) criteria (see Table 1). Only four studies received scores above 6. These four studies used appropriate comparison groups and three utilized a reliable and valid instrument to code key elements in the health professional’s capacity report. Accordingly, these studies received strong (Moye et al., 2007b) and moderately strong (Keith and Wacker, 1993; Dudley and Goins, 2003; Gibson, 2011) classifications based on their methodological quality. A strength of Keith and Wacker’s (1993) study was the use of four comparison groups that were sourced from two time periods, before and after law reform, allowing testing of hypotheses that were based on change over time. In contrast, the seven articles that were classified as weak were each based on a single group design, lacked comparative statistical analysis, and did not address the effect of any design limitations on the findings (Peters et al., 1985; Bulcroft et al., 1991; Stevenson and Capezuti, 1991; Juretic et al., 1993; Barritt Lisi and BarinagaBurch, 1995; Parker, 2008; Björkstén et al., 2014) (Table 2). DATA EXTRACTION METHODS
The same three studies that utilized a reliable and valid instrument, systematically coded key elements from capacity reports (Dudley and Goins, 2003; Moye et al., 2007b; Gibson, 2011). Two studies developed survey instruments with high interrater reliability; Juretic and colleagues (1993) used a 45 item measure based on expert consensus (99% agreement) and Bulcroft and colleagues (1991) developed a semi-structured instrument of 40 items (Kappa values of reliability ranged from .85 to .98). Barritt Lisi and Barinaga-Burch (1995) employed two survey instruments, one for use when observing the proceedings, the other for a file summary. Keith and Wacker (1993) counted the frequency of clinical reports contained in guardianship files before and after legislative changes and Björkstén et al. (2014) counted the frequency of responses contained in physician statement forms issued for the guardianship court. Three articles did not describe their data extraction methods (Peters et al., 1985; Stevenson and Capezuti, 1991; Parker, 2008). Study findings SUBJECTS OF GUARDIANSHIP APPLICATIONS
The majority of individuals were elderly, with the reported average age across studies ranging from 76 years to 82 years. Most were female and
379
unmarried, including those who were widowed or divorced, and they often resided in hospitals, nursing homes or residential facilities at the time of the guardianship hearing. Three of the studies reported the majority of persons were diagnosed with a dementia, ranging from 59% to 71% across the samples, followed by a mental illness (20–24%) and other neurological events such as stroke (11– 24%) (Peters et al., 1985; Stevenson and Capezuti, 1991; Moye et al., 2007b). Five studies recorded that persons appeared at their own hearings between
REVIEW
Decision-making capacity evaluation in adult guardianship: a systematic review ...........................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................
Sally McSwiggan, Susanne Meares and Melanie Porter Department of Psychology, Macquarie University, North Ryde, Australia
ABSTRACT
Background: Evidence of impaired decision-making capacity is a legal requirement for adult guardianship. To understand the quality of the evidence health professionals commonly provide in reports submitted to guardianship courts, a systematic review was undertaken to appraise the design and methodological quality of the published literature on health professionals’ written reports of decision-making capacity and to describe the content of these reports. Methods: Electronic searches from 1980 to 2015 identified 1183 articles of which 11 met the inclusion criteria where each evaluated quantitatively the content of health professionals’ written reports submitted to adult guardianship proceedings. Methodological quality of the selected studies was rated using a critical appraisal tool. Results: Nine studies sourced files from courts in the U.S. and one from Sweden; another reported on guardianship decisions from Australia. Four studies were rated as moderately strong or strong. Strengths included the use of comparison groups and a reliable and valid instrument to code reports. The review showed a person’s medical condition was often cited as evidence of impaired decision-making capacity. Cognitive, psychiatric and functional abilities were less often described, and a person’s values and preferences were rarely recorded. Conclusions: It is recommended health professionals describe the process by which a person makes a particular decision (their ability to understand, appreciate, reason and communicate) in addition to providing medical information, including cognitive, psychiatric and functional abilities. This approach provides support for a professional’s opinion and evidence for a court. International studies of health professionals’ approach to decision-making capacity evaluation are needed. Key words: decision-making capacity, competency, adult guardianship, systematic review
Introduction Capacity refers to a judgement as to whether an individual has the requisite cognitive, decisional, affective, and practical skills to be found to have the ability to adequately complete a particular task (such as driving), or to make a particular decision (such as to refuse medical treatment) (Moye and Braun, 2010). Decisional abilities are distinct from practical (also known as executional) abilities in so far as the person may not be able to accomplish Correspondence should be addressed to: Sally McSwiggan, Institute for Biomedical Ethics, Basel University, Bernoullistrasse 28, Basel 4056, Switzerland. Phone: +41 (0) 61 267 17 86, +41 (0) 79 244 75 78; Fax +41 (0) 61 267 17 80. Email: [email protected]. Received 24 Mar 2015; revision requested 13 Jul 2015; revised version received 17 Aug 2015; accepted 18 Aug 2015. First published online 28 September 2015.
their choice without assistance (Collopy, 1995; Boyle, 2008). Although conclusions about capacity are ultimately determined by the law, in practice most views formed about a person’s capacity are established by health professionals, lawyers, social care workers, friends and family (Kapp, 2002; Bennett and Hallen, 2005). When a view is formed that a person has impaired capacity, the vast majority of solutions are informal such as family or friends offering assistance. However, when informal solutions cannot be found or issues of conflict are raised, the law may have a role. Applications are made to guardianship (also known as conservatorship) courts and tribunals seeking legal orders to appoint a substitute decisionmaker, also known as a guardian or conservator (Leatherman and Goethe, 2009). A critical piece
374
S. McSwiggan et al.
of evidence in the legal proceedings is the clinical report of decision-making capacity written by a health professional (Moye et al., 2005). Models of decision-making capacity that have developed over the last 30 years have been based on empirical research, expert consensus and case law (Appelbaum and Roth, 1981; Grisso and Appelbaum, 1998; Moye and Marson, 2007; Karlawish, 2008). Key components include: (1) the presumption of capacity, (2) capacity as domain specific to the lifestyle area under review (e.g., finances or personal care), (3) decision specific within that domain (e.g., paying bills or selling real estate), and (4) temporally limited due to the transient and progressive nature of some disabilities (e.g., delirium or Alzheimer’s disease) (Ganzini et al., 2003; Moberg and Kniele, 2006; Boyle, 2008; Newberry and Pachet, 2008). Research has helped to advance the understanding that decision-making capacity is not a global, “all or nothing” status but rather a situation-specific functional definition (Anderer, 1990; Sabatino and Basinger, 2000; Peisah et al., 2009). Reviews of empirical research grouped by cognitive and psychiatric disorders have reliably found that medical diagnoses alone do not predict impaired decision-making capacity (Okai et al., 2007; Moye et al., 2006; van Staden, 2009). More specifically, no single neuropsychological test, cognitive domain or psychiatric presentation consistently predicts a lack of capacity (Okai et al., 2007; Palmer and Savla, 2007; Moberg and Rick, 2008). Additionally, mental status screening has little utility beyond staging cognitive impairment (Pachet et al., 2010). The results show that decision-making capacity is more than group membership or scores achieved on standardized tests. Surveys of health professionals have suggested decision-making capacity is often considered challenging to assess and report, commonly due to limited knowledge of the concept (Ganzini et al., 2003; Hill et al., 2006; Mullaly et al., 2007). To aid health professionals in providing clinical reports of decision-making capacity for adult guardianship courts and tribunals, there has been an emergence of resources (e.g., Ontario Ministry of the Attorney General, 2005; Moye et al., 2007a; American Bar Association Commission on Law and Aging and American Psychological Association, 2008; Australian Psychological Society and the Office of the Public Advocate, 2009; British Psychological Society, 2010; O’Neill and Peisah, 2011). To date, however, no systematic review has been undertaken of the literature that has examined the content of written reports of decisionmaking capacity tendered by health professionals for adult guardianship proceedings. The aims of this systematic review were (a) to critically appraise
the design and methodological quality of the published literature on decision-making capacity reports written by health professionals for adult guardianship proceedings and (b) to describe the content of these reports.
Methods Searches were conducted using three electronic databases: Ovid MEDLINE (1948 to February, 2015), PsycINFO (1967 to February, 2015) and Scopus (1980 to February, 2015) from 1980 until February 1, 2015 (Figure 1).The search terms were developed based on specific key words to capture guardianship proceedings (guardianship/ guardian/conservatorship/conservator) and capacity (capacity/competency/decision making/decision making capacity/incompetent/incompetence/incompetency/mental capacity/mental incapacity) or disability (disability/dementia/brain injury/cognition/ cognitive impairment) or proceedings (proceedings/determination/order/finding/evidence). The selection criteria for inclusion in the systematic review were as follows: (a) the article was written in English, (b) participants were aged 18 years or above, (c) capacity reports for guardianship proceedings were evaluated and (d) the article reported results using quantitative analyses. The exclusion criteria were as follows: (a) opinion articles and narrative reviews, (b) studies containing less than 10 subjects. Articles meeting the selection criteria were rated for methodological quality in 11 key areas using a 14-item critical appraisal tool developed by Heacock et al. (1997; Table 1). Scores ranged from 0–12. Heacock and colleagues (1997) categorized scores as strong (9–12 points), moderately strong (more than 6 and up to 9 points) and weak (6 points and less). The critical appraisal tool was developed for non-experimental quantitative research. It was selected for use in the current study because a guide to scoring each item was provided, and the scale had undergone content validation and reliability testing (Crowe and Sheppard, 2011). Heacock and colleagues (1997) indicate good to excellent interrater reliability of the checklist items apart from the first item which was reported as poor. Each of the key areas was weighted equally in terms of importance for methodological quality and was assigned a value of one point in the checklist with one exception being area 9 (Appropriateness of Statistical Test/Sample Size) that was assigned a value of two because it was considered vital to the valid interpretation of research results (Heacock et al., 1997). This weighting scheme attempts to address the problem whereby summary scores hide defects if studies score well in some areas considered
Decision-making capacity in adult guardianship
375
Figure 1. Flowchart of the search strategy.
not as methodologically important as others (Crowe and Sheppard, 2011). Results of the database searches were downloaded and imported into Endnote files. Duplicate articles were excluded. The title and abstracts of all articles were reviewed by the first author to identify those meeting selection criteria. Where abstracts were not available or study methods were ambiguous, the full text of the article was sourced and examined. The full texts of all articles meeting selection criteria were retrieved. Reference lists of the articles meeting the selection criteria, opinion papers and narrative reviews were searched for additional relevant articles.
Results The results of the searches produced 961 articles from Ovid MEDLINE and PsycINFO and 222 from Scopus. Ten articles met the inclusion criteria. Another article was sourced by examining the references from the selected articles, bringing the total to eleven. Figure 1 provides details of the identification process, including the reasons for exclusion.
Study quality Research questions examining reports of capacity written for guardianship proceedings are restricted in so far as experimental designs cannot be applied due to the ethical and legal constraints imposed by the lawful process being examined. These constraints limit research designs to observational studies which are largely descriptive. Of the articles selected, ten used retrospective cross-sectional designs and one was a before and after design (Keith and Wacker, 1993). Capacity reports were generally sourced from guardianship files held by the courts. As shown in Table 2, nine studies sourced files from between one and 30 courts across various states in the U.S. and one other sourced files from a court in Sweden. Reports were variously selected by randomizing guardianship files (Stevenson and Capezuti, 1991), choosing alphabetically-consecutive files to meet a quota (Bulcroft et al., 1991), convenience samples (Jurectic et al., 1993) and every other consecutive application (Björkstén et al., 2014). The ten studies examined in total 2443 guardianship files from the U.S. and Sweden. The remaining study reported on 285 reasons for decision (also known as judgements) from guardianship hearings in Australia (see Table 2).
376
BARRITT STEVENSON KEITH
KEY AREAS
ITEMS
LISI AND
DUDLEY
JURETIC BARINAGA- AND
MOYE
BJÖRKSTÉN
PETERS BULCROFT AND
AND
ET AL., ET AL.,
CAPEZUTI,
WACKER, ET AL.,
BURCH,
GOINS,
ET AL., PARKER, GIBSON, ET AL.,
1985
1991
1993
1995
2003
2007
1991
1993
2008
2011
2014
...........................................................................................................................................................................................................................................................................................................................................................................................................................................................
Gap in literature Research question Methodology Measurement Representativeness Representativeness Control/comparison group Sample size/power Statistical test(s) Limitations/bias Limitations/bias Statistical test/sample size Implicationsa Implicationsa
1 2 3 4 5 6 7 8 9 10 11 12 13 14 Total
1 1 0 0 0.5 0.5 0 0 0 0.5 0 2 0 0 5.5
1 1 1 0 0.5 0.5 0 0 0 0 0 2 0 0 6
0 1 1 0 0.5 0.5 0 0 0 0 0 2 0 0 5
1 1 1 0 0.5 0.5 1 0 1 0 0 2 0 0 8
0 1 1 0 0.5 0.5 0 0 0 0 0 2 0 0 5
1 1 0 0 0.5 0.5 0 0 0 0 0 2 0 0 5
1 1 1 1 0.5 0.5 1 0 0.5 0 0 2 0 0 8.5
1 1 1 1 0.5 0.5 1 0 1 0.5 0.5 2 0.5 0.5 11
0 1 1 0 0.5 0.5 0 0 0 0 0 2 0 0 5
0 1 1 1 0.5 0.5 1 0 1 0 0 2 0 0 8
1 1 1 0 0.5 0.5 0 0 0 0 0 2 0 0 6
Notes: Gap in Literature: 1 = Does the study identify a gap in the existing literature; Research Question: 2 = Is the research question clearly stated; Methodology: 3 = After reading the methods section, could the reader repeat the study with confidence; Measurement Instruments/ Techniques: 4 = Do the authors state the measurement technique(s)/ instrument(s) are valid/ reliable or do they provide information so the reader can investigate the validity and reliability of the measurement technique(s)/ instrument(s); Representativeness: 5 = Is the study population representative of the end-user population; Representativeness: 6 = Is the study setting representative of the workplace to which the results will be applied; Control/ Comparison Group: 7 = Is there an appropriate control/comparison group; Sample Size: 8 = Is there a statement about sample size or power of the study; Statistical Test(s): 9 = Do the authors state the statistical test used and level of significance or confidence levels achieved; Limitations/Bias: 10 = Do the authors discuss the limitations or biases of the study design and/or methodology; Limitations/ Bias: 11 = Do the authors discuss the effects of the preceding limitations or biases on the results; Appropriateness of Statistical Test/ Sample Size: 12 = Do the authors achieve any one of the following; utilize a study sample with at least 30 subjects or; conduct non-parametric statistics tests for samples with less than 30 subjects or; conduct parametric tests for samples with less than 30 subjects but explain the sample is normally distributed or; conduct multivariate tests with ten times as many subjects as there are independent variables; Implications: 13 = Do the authors discuss recommendations of the findings; Implications:14 = Do the authors identify opportunities for further research based on the current study. Methodological ratings are from Heacock et al., (1997). a Only answer items 13 and 14 if you responded yes to item 11.
S. McSwiggan et al.
Table 1. Critical appraisal ratings of the selected articles
Table 2. Data extraction from the selected articles ARTICLE (CAT SCORE) ( C A T C L A S S I FI C A T I O N )
SAMPLE
METHOD
LOCATION
STATISTCAL ANALYSIS
RESULTS
...........................................................................................................................................................................................................................................................................................................................................................................................................................................................
N = 42 cases aged ࣙ 18 years
Bulcroft et al. (1991) (6/12) (weak)
N = 63 cases aged ࣙ 50 years
Stevenson and Capezuti (1991) (5/12) (weak)
N = 49 cases aged ࣙ 60 years
Keith and Wacker (1993) (8/12) (moderately strong)
N = 766 cases aged ࣙ 60 years exclusion: voluntary guardianship N = 40 cases aged ࣙ 60 years exclusion: acute trauma
Barritt Lisi and Barinaga-Burch (1995) (5/12) (weak)
Dudley and Goins (2003) (8.5/12) (moderately strong)
Juretic et al. (1993) (5/12) (weak)
Written reports and 8 observed hearings from one court, January, 1977 to May, 1982 Written reports from two courts, alphabeticallyconsecutive files were selected to meet a quota, January, 1987 to December, 1988 Written reports from one court; reports were selected at random, 1987
FL, U.S.
frequency
7% described symptoms of physical incapacity; 12% described symptoms of mental incapacity
WA, OH, U.S.
frequency interrater reliability
Demonstration of incompetence “rarely supplied”; most common phrase “failure to manage estate,” followed by “confusion/ disorientation”
PA, U.S.
frequency
Written reports from court files before/after legislation reform, 1984 (IA), 1983 (MO)
IA, MO, U.S.
frequency t-test
>50% medical diagnosis only evidence; 20% nonspecific diagnosis “organic brain syndrome”; 10% “senile” 14% increase in “medical conditions” described (IA) after legislative reform
Written reports from four courts selected on availability, June, 1988 to June, 1989
TX, U.S.
frequency interrater reliability
N = 726 cases aged ࣙ 60 years
Written reports and observed hearings from 30 courts, October, 1989 to January, 1992
frequency
N = 119 cases aged ࣙ 65 years exclusion: reviews of guardianship and petitions for conservatorship
Written reports from six courts, December, 1996 to December, 1999
CA, CO, FL, IN, KS, MI, MN, NY, OR, WA, U.S. PA, WV, U.S.
frequency chi-square
75% medical or psychiatric diagnosis; 8% mental status examination; 48% unsupported statements of incompetence; 8% prognosis or potential need for re-evaluation >50% medical evidence
95% (PA) versus 66% (WV) cognitive assessment; 2% overall met minimum reporting criteria in all clinical areas (cognitive, psychiatric, functional, current treatment)
Decision-making capacity in adult guardianship
Peters et al. (1985) (5.5/12) (weak)
377
378
ARTICLE (CAT SCORE) ( C A T C L A S S I FI C A T I O N )
SAMPLE
METHOD
LOCATION
STATISTCAL ANALYSIS
RESULTS
...........................................................................................................................................................................................................................................................................................................................................................................................................................................................
Moye et al. (2007b) (11/12) (strong)
N = 298 cases aged ࣙ 55 years exclusion: developmental delay
Written reports and oral transcriptions from eight courts, January, 2002 to December, 2005
MA, PA, CO, U.S.
frequency chi-square
Parker (2008) (5/12) (weak)
N = 285 reasons for decision
QLD, Australia
frequency
Gibson (2011) (8/12) (moderately strong)
N = 80 cases (240 capacity reports) aged ࣙ 50 years
Published reasons for decision, AUSTLII website, January, 2005 to April, 2008 Written reports and oral transcriptions from two courts, 2004, 2005, 2006
KY, U.S.
frequency chi-square one-way analysis of variance
Björkstén et al. (2014) (6/12) (weak)
N = 260 cases aged ࣙ 50 years exclusion: involuntary guardianship and normal memory
STK, Sweden
frequency
Written reports from one court; every other consecutive application was selected, June 2007 to July, 2008
86%, 63%, 87% medical diagnosis (MA, PA, CO); 61%, 47%, 86% cognitive abilities (MA, PA, CO); 5%, 18%, 23% mood disturbance (MA, PA, CO); 5%, 5%, 19% cognitive screening (MA, PA, CO); 1%, 2%, 34% neuropsychological assessment (MA, PA, CO) 76% of reasons contained legal findings on capacity;44% had evidence from two or more experts; 29% disagreed 88%, 95%, 94% medical diagnosis; 53%, 91%, 89% cognitive abilities; 23%, 71%, 77% functional abilities (physician, psychologist, social worker) 20% mental status testing or neuropsychological assessment
Notes: CAT = critical appraisal tool; CA = California; CO = Colorado; IA = Iowa; IN = Indiana; FL = Florida; KS = Kansas; KY = Kentucky; MA = Massachusetts; MI = Michigan; MN = Minnesota; MO = Missouri; NY = New York; OH = Ohio; OR = Oregon; QLD = Queensland; STK = Stockholm; TX = Texas; WA = Washington; WV = West Virginia.
S. McSwiggan et al.
Table 2. Continued.
Decision-making capacity in adult guardianship
The methodological rating scores for the eleven articles ranged from 5 to 11 according to the Heacock et al. (1997) criteria (see Table 1). Only four studies received scores above 6. These four studies used appropriate comparison groups and three utilized a reliable and valid instrument to code key elements in the health professional’s capacity report. Accordingly, these studies received strong (Moye et al., 2007b) and moderately strong (Keith and Wacker, 1993; Dudley and Goins, 2003; Gibson, 2011) classifications based on their methodological quality. A strength of Keith and Wacker’s (1993) study was the use of four comparison groups that were sourced from two time periods, before and after law reform, allowing testing of hypotheses that were based on change over time. In contrast, the seven articles that were classified as weak were each based on a single group design, lacked comparative statistical analysis, and did not address the effect of any design limitations on the findings (Peters et al., 1985; Bulcroft et al., 1991; Stevenson and Capezuti, 1991; Juretic et al., 1993; Barritt Lisi and BarinagaBurch, 1995; Parker, 2008; Björkstén et al., 2014) (Table 2). DATA EXTRACTION METHODS
The same three studies that utilized a reliable and valid instrument, systematically coded key elements from capacity reports (Dudley and Goins, 2003; Moye et al., 2007b; Gibson, 2011). Two studies developed survey instruments with high interrater reliability; Juretic and colleagues (1993) used a 45 item measure based on expert consensus (99% agreement) and Bulcroft and colleagues (1991) developed a semi-structured instrument of 40 items (Kappa values of reliability ranged from .85 to .98). Barritt Lisi and Barinaga-Burch (1995) employed two survey instruments, one for use when observing the proceedings, the other for a file summary. Keith and Wacker (1993) counted the frequency of clinical reports contained in guardianship files before and after legislative changes and Björkstén et al. (2014) counted the frequency of responses contained in physician statement forms issued for the guardianship court. Three articles did not describe their data extraction methods (Peters et al., 1985; Stevenson and Capezuti, 1991; Parker, 2008). Study findings SUBJECTS OF GUARDIANSHIP APPLICATIONS
The majority of individuals were elderly, with the reported average age across studies ranging from 76 years to 82 years. Most were female and
379
unmarried, including those who were widowed or divorced, and they often resided in hospitals, nursing homes or residential facilities at the time of the guardianship hearing. Three of the studies reported the majority of persons were diagnosed with a dementia, ranging from 59% to 71% across the samples, followed by a mental illness (20–24%) and other neurological events such as stroke (11– 24%) (Peters et al., 1985; Stevenson and Capezuti, 1991; Moye et al., 2007b). Five studies recorded that persons appeared at their own hearings between
