Adequacy of measures of informed consent in medical practice: A systematic review

As a critical component of medical practice, it is alarming that patient informed consent does not always reflect (1) adequate information provision, (2) comprehension of provided information, and (3) a voluntary decision. Consequences of poor informed consent include low patient satisfaction, compromised treatment adherence, and litigation against medical practitioners. To ensure a well-informed, well-comprehended, and voluntary consent process, the objective and replicable measurement of these domains via psychometrically sound self-report measures is critical. This systematic review aimed to evaluate the adequacy of existing measures in terms of the extent to which they assess the three domains of informed consent, are psychometrically sound and acceptable for use by patients. Extensive searching of multiple databases (PsychINFO, PubMed, Sociological Abstracts, CINAHL, AMED) yielded 10,000 potential studies, with 16 relevant scales identified. No existing scale was found to measure all three consent domains, with most only narrowly assessing aspects of any one domain. Information provision was the most frequently assessed domain, followed by comprehension, and then voluntariness. None of the identified scales were found to have adequate evidence for either high quality psychometric properties or patient user acceptability. No existing scale is fit for purpose in comprehensively assessing all domains of informed consent. In the absence of any existing measure meeting the necessary criteria relating to information, comprehension and voluntariness, there is an urgent need for a new measure of medical consent to be developed that is psychometrically sound, spans all three domains and is acceptable to patients and clinicians alike. These findings provide the impetus and justification for the redesign of the informed consent process, with the aim to provide a robust, reliable and replicable process that will in turn improve the quality of the patient experience and care provided.


Introduction
Obtaining patient consent is a necessary and critical process enshrined in medical practice, that is characterised within a model of three domains: (provision of) Information, Comprehension (of information by the patient), and Voluntariness (of the patient's decision without coercion) [1]. The complexity and unfamiliarity of many medical decisions renders decision making challenging for patients with many rarely involved in a shared decision-making process [2,3]; yet health professionals do not necessarily have a good grasp of the consent process and patient decision-making capacity [4]. Although aspects of consent may be fulfilled by patients completing a consent form [5,6], research indicates the informed consent process is poorly executed, with widespread reporting of ill-informed patients who agree to procedures for which they have little comprehension [7], suggesting that the provision of consent does not always reflect a well-informed, well-considered, ethical process [5]. With shared decision making the gold standard for medical practice [8][9][10], it is critical that the medical consent process reflects the patient's values and goals, and not simply the clinician's view [11]. There are costly consequences of patients being inadequately informed about medical procedures ranging from low patient satisfaction with care, increased patient regret, poor adherence to treatment plans, both underuse and overuse of the health system [2,[12][13][14][15][16][17], and patient litigation against medical practitioners [18]. These failures are due in part to the lack of any standardised approach to measuring the consent process [9], and there is a pressing need for standardisation across medical practice contexts [19]. Reviews of the informed consent process for participation in medical research and clinical trials similarly highlight serious shortcomings stemming from lack of standardisation of measures [20][21][22]. To ensure that truly informed consent is achieved, it is critical that the patient perspective is reflected by assessing the three consent domains using patient-reported measures undertaken in an objective and replicable manner (i.e., using psychometrically-sound measures). Yet, no such systematic investigation has been undertaken to evaluate the quality of current measures available to assess the informed consent process in medical practice, or whether current measures are fit-for-purpose in assessing the three consent domains. Adequately assessing the extent to which the three domains of consent are being addressed requires clarification of the breadth of these constructs. Providing information, the first critical step in the consent process, entails describing the procedure, anticipated risks/benefits, and available alternatives [3], as well as facilitating the patient to ask questions beyond that which has been provided in the consent documentation [23]. However, there is frequently a mismatch between information given by health professionals and that desired by patients [24,25].
Closely related to the need for appropriate provision of information is the requirement for the patient to comprehend the information, including their ability to apply the information to their own situation, using this to support their decision [26]. The temporal aspect of comprehension is also important to assess in terms of whether a patient has had sufficient time to make a decision [2,21,27]. Comprehension further entails consideration of whether the individual has the cognitive capacity to fully process and understand the information provided [19,28,29], suggesting the need to evaluate consent capacity [30] and tailor consent information that accommodates for differing cognitive abilities (e.g., graphical and simplified textbased formats) [31,32]. There is documented widespread lack of comprehension [22], particularly regarding patient understanding of benefits/risks of surgery and anaesthesia, and alternative treatment options [5,[32][33][34]. Clearly, any assessment of medical consent should comprehensively capture both the extent to which a patient is informed, as well as the different dimension reflecting their comprehension.
Other factors that may influence informed consent decision-making include cultural or language barriers, and fear of the healthcare system, that is frequently present in minority populations [e.g., 28,35,36]. This relates to the third domain of voluntariness, that stipulates the need for informed consent to be undertaken without any duress or undue influence from medical practitioners. Of note, however, is research highlighting that the level of independence in medical decision making differs from person to person, with some individuals requiring greater assistance in arriving at a voluntary decision, whereas others require more time to think and consider the decision on their own. Another critical aspect of voluntariness is the assessment of comprehension and information provision independent of the clinician through patient self-reported measures, whose direct questions through clinical interview may introduce bias by influencing responder's answers, negating the voluntariness of the decision [19].
An objective and replicable approach is needed to measure the extent to which consent processes are meeting the goals of the three domains of Information, Comprehension, and Voluntariness. Prior systematic reviews have not directly addressed this issue, but relatedly, have highlighted a paucity of reliable and valid scales to assess specific aspects involved in medical decision-making more generally (e.g., extent of patient involvement in decision making [9], patient need for information [25], and participation in clinical trials [20]). To that end, a detailed examination of the measures available to assess the consent process in medical practice is warranted. This study aimed to undertake a systematic review to identify and critically evaluate existing evidence-based measures of patient informed medical consent. Specifically, we aimed to evaluate the adequacy of existing measures in terms of the extent to which they assess the three domains of informed consent and whether they are psychometrically sound, reliable and valid measures. This review paper will make recommendations for the potential application or modification of relevant existing measures and identify gaps in currently available measures for assessing medical consent, in order to provide a robust, reliable and replicable process.

Methods
This systematic review was conducted in accordance with the PRISMA guidelines (see S1 Checklist for PRISMA checklist).

Search strategy
Six online journal databases (PsychINFO, PubMed, Sociological Abstracts, CINAHL, AMED) were searched for relevant articles in March 2020 using the following detailed search strategy: (consent � OR withdraw � ) OR ((decision � AND (making OR capac � OR conflict � OR regret � OR need � OR support � )) OR competence OR risk � OR coer � OR "mental capac � ") OR (underst � OR comprehen � OR benefit � OR harm � OR literacy OR know � OR aware � OR (satisfac � AND (patient OR client)) OR disclos � OR (treat � AND (option � OR alternati � ))) OR (autono � OR agen � OR self-determin � OR value � OR choice) OR ("physician-patient relations" OR communication OR dissemination OR "privilege � communic � " OR care OR services OR treatment OR privacy) OR (clinical AND trial OR randomi?ed AND control?ed AND trial OR rct) AND (psychomet � OR valida � OR measure � OR scal � OR survey � OR questionnaire � OR (construction AND (test OR scale))). Reference lists of identified articles were hand searched to ensure all potential studies were captured. The review protocol is not published elsewhere.

Inclusion and exclusion criteria
The review inclusion and exclusion criteria are listed in Table 1. Wherever the abstract or title indicated exclusion criteria were met, the study was excluded. Full text articles were reviewed when it was unclear if inclusion criteria were met from the title or abstract alone and discussed amongst the research team; papers that met all the inclusion criteria and which did not meet any of the exclusion criteria were included for review.

Scale evaluation scoring system
The authors reviewed the included studies for evidence of scale validation. Scale psychometric properties were rated using a previously published scoring system [37], see Table 2. A score of "1" was given when the validation sample exceeded 300, and "0.5" for sample sizes >200 and <300 [38]. For every reported adequate psychometric property (e.g., construct validity) [39], studies received a score of "1" or "0.5" (depending on the scoring of each psychometric property) [40]. When authors did not examine a type of validation or the validation failed, it was scored "0". Scores for each psychometric property were summed to give an overall score (possible range: 0-16). Ratings were undertaken by two of the authors (MP, BS) with any disagreements discussed until agreement was reached in collaboration with the other authors (KAS, CJK).

Results
The results of the literature search are presented in the PRISMA diagram (Fig 1). From 10,663 articles initially identified, 10,370 were unique results (n = 293 removed), and after screening for titles and abstracts, a further 10,304 articles were excluded. Of the 66 remaining articles, 8 met inclusion criteria. Hand-searching of reference lists of these included articles yielded a further 8 papers that met inclusion criteria, resulting in 16 included articles (representing 19 different studies reporting on 16 different scales).

PLOS ONE
Adequacy of medical practice informed consent measures: A systematic review   (1 = strongly disagree to 5 = strongly agree).
Validity. None of the scales were awarded a full score (i.e., 10) for evidence supporting their validity. Of those with the greatest evidence for validity, COMRADE [41] scored 6/10 and [47, 51, 54] each scored 5/10. Acceptability to users. Only one scale provided information from the patient perspective on acceptability of the scale, specifically regarding comfort in responding to the questions (Brief Health Literacy) [54]. Information about completion rates was provided by 5 (38%) scales: [41,42,46,49,50].

PLOS ONE
Adequacy of medical practice informed consent measures: A systematic review

Discussion
This systematic review identified and evaluated existing approaches to measuring the adequacy of the informed consent process in medical practice. After an extensive literature search of more than 10,000 potential studies, only 16 scales were identified that assess any of the three domains of informed consent. Most of the identified scales have extensive limitations. In terms of the content that is being assessed, no existing scales measure all three consent domains. Several scales measure two of the three domains [41,42,[45][46][47]53]. Moreover, the exact aspect of each domain tapped into by these measures were quite different. Hence, in terms of content being assessed, no one scale is fit for purpose in comprehensively assessing all domains of informed consent, and at best several measures would need to be used in conjunction to assess all domains comprehensively.
Of the three top ranking scales assessing the Information domain, the COMRADE [41] focuses solely on whether risk-related information has been provided for the patient, and the Facilitation of Patient Involvement Scale [43] and the HCEQ [47] capture the extent to which   the patient perceives that the physician has facilitated a shared decision. Of the two latter scales, the Facilitation of Patient Involvement Scale [43] has more extensive evidence for validity (criterion and construct) having demonstrated that it can predict patient satisfaction following medical procedures and demonstrates logical relationships with similar constructs. As such, a combination of components from the COMRADE [41] and the Facilitation of Patient Involvement Scale [43] may be appropriate to comprehensively assess the Information domain.
The three highest scoring scales for the domain of Comprehension [44,46,54] have similar levels of evidence for their validity, with one scale [44] additionally providing evidence for stability over time. The Decisional Conflict Scale [44] captures extent of uncertainty around decision making, whereas the Brief Health Literacy Screen [54] assesses self-perceptions of comprehending the information provided by a physician. More directly related to the informed consent process is the informed choice subscale of the Decision Evaluation Scales [46], assessing self-perceptions of being adequately informed about a procedure in order to make a choice. Perhaps due to the inherent difficulty in developing measures that assess context-specific knowledge (e.g., details of a surgical procedure), none of these scales assess accuracy of knowledge. What is missing from any of these higher rated scales is consideration of whether the patient is able to apply the information provided to their own situation, using this to support their decision making [26], and the temporal aspect of the consenting process [2,27], two factors that are key to comprehension.
For the Voluntariness domain, the COMRADE [41] scored highest with a subscale that assesses patient confidence to make a decision, followed by the API [45] and HCEQ [47] both of which assess the patient's preferred role or amount of control in the decision-making process. The API [45] has greater evidence for reliability and validity, compared with the HCEQ [47], potentially making it the more useful measure. Although tapping into aspects of patient confidence and patient preferred role in this decision making, an aspect neglected by these higher rated scales is whether the consent (decision) is being made free of undue influence or duress from the medical team, the main caveat of voluntariness [19].
In assessing the quality of the existing scales reflecting aspects of the informed consent process, it is important to consider how the scale may be implemented in clinical practice in terms of: (1) does the scale have high quality psychometric properties to ensure that the variability in scores on the scale truly reflect variability in the underlying construct (e.g., comprehension), rather than measurement error?; (2) does the scale preferably measure more than one domain of informed consent, to minimize patient burden and have the most parsimonious measurement approach?; and, (3) practical considerations regarding user (patient) experience of using the scale, and the administration, scoring and interpretation of the scale. For psychometric quality, the COMRADE [41] scored highest, followed by [43-45, 47, 54], with three of these scales assessing more than one domain [41,45,47]. Information is the most frequently assessed domain of these higher quality scales, yet, the COMRADE assesses a very narrow perspective (i.e., whether risk-related information has been provided), neglecting crucial aspects including whether the procedure and available alternatives are adequately described [3] and if the patient was able to ask questions prior to consenting [23]. Hence, as a measure of the Information domain, COMRADE alone is not fit for purpose and needs to be supplemented. In the interests of providing a parsimonious approach to measuring informed consent, the COM-RADE does provide a measure of Voluntariness, although this alone is not fit for purpose as it neglects to assess whether consent (the decision) is being made free of undue influence or duress from the medical team [19]. Moreover, since the COMRADE neglects to assess comprehension, other high scoring measures in this domain (e.g., Brief Health Literacy Screen [54]) may need to be invoked. Importantly, in reviewing the 16 scales assessing aspects of the consent process, only the San Diego Brief Assessment of Capacity to Consent [56] was purposely designed for measuring the consent process, specifically in randomized controlled trial research. However, this scale has yet to be subjected to rigorous validity assessment. All other measures were designed to assess specific aspects of communication or decision making relating to medical procedures but were not purpose built to comprehensively measure any one of the domains of consent, calling into question the adequacy of even a combination of these measures to assess each domain.
The results of this review should be considered with some key limitations in mind. As is the case with any review, the adopted search strategy may not have identified all relevant articles. Yet, our rigorous broad search strategy and extensive manual review of the large volume of initial papers support the notion that the results of this review are representative of the field. Further, the review did not look at the consent process for individuals who cannot give consent for themselves (e.g., issues of soundness of mind or legalities in consent such as in minors). In addition, this review did not consider the impact of culture on medical consent, which may influence the level of voluntariness for example, and thus, culture-specific scales may be needed in certain collectivistic countries to adequately assess medical consent. These were issues beyond the scope of this review and were therefore not included. However, future work in the medical consent field may benefit from a similar review as this one on these more complex situations. Lastly, this review did not include qualitative studies of medical consent scales and hence did not allow for a more in-depth analysis of the patients' subjective experience with the scales; it is recommended that future research elicit patients' responses to open-ended questions about their experiences with medical consent scales.

Conclusion
This review has highlighted the complex nature of consent reflected within the three domains of information, comprehension and voluntariness, and the need to ensure that these multiple aspects are reflected in both measurement and practice. Of greatest concern is the absence of any established, standardized, reliable and valid measure of consent addressing these three domains, suggesting serious limitations in the current approaches to the consenting process. In summary, as a starting point, the COMRADE is the most likely contender for assessing certain aspects of informed consent. Yet, it only provides a modest level of coverage of the informed consent domains, and it lacks clear evidence for validity and user acceptabilitysomething inherently lacking in most of the reviewed scales. Clearly, what is needed is not only a psychometrically high-quality measure, but also one that is easy for patients to use and has a high degree of perceived relevance to patients and clinicians alike. Unfortunately, only one of the higher scoring scales reviewed, the Brief Health Literacy Screen, assessed the usability of the scale, yet this scale only covers one of the elements of informed consent (Comprehensiveness). Therefore, there is need for a new measure of medical consent to be developed that has robust psychometric properties, spans all three domains and is acceptable to patients and clinicians alike, that will in turn improve the quality of the patient experience and care provided.