Measuring Trust in Primary Care

Abstract

Policy Points.

• Trust in primary care clinicians is essential for effective patient care and is associated with better health outcomes, but it is rarely assessed, and existing measures have not been thoroughly evaluated.

• This scoping review reveals that research assessing patients’ trust in primary care clinicians largely stopped more than a decade ago but offers candidate measures for future testing, implementation, and policy applications.

Context

Trust is a fundamental aspect of any human relationship, and medical care is no exception. An ongoing, trusting relationship between clinicians and patients has shown demonstrable value to primary care. However, there is currently no measure of trust in general use, and none endorsed for use by most value‐based payment programs. This review searched the literature for any existing measures of patient trust in primary care clinicians and assessed their potential to be implemented as a patient‐reported outcome measure.

Methods

A keyword search on PubMed along with scanning references was conducted to find any trust measures in health care. Measures that did not address primary care clinicians were eliminated and the remaining measures were then assessed for their utility to primary care.

Results

This purposeful, scoping review found four tested measures for assessing patients’ trust in primary care clinicians that are candidates for general use. Of these four, the revised Trust in Physicians Scale and Wake Forest Physician Trust Scale are the most tested and viable options.

Conclusion

Renewed national interest in trust in health care should focus on the capacity to measure it. This review informs the effort to test trust measures for use in research, practice improvement, and value‐based payment. Measuring trust, how it relates to outcomes, and learning how it is produced or lost are key to assisting practices and health systems toward earning it.

In a 2021 white house virtual town hall, white house Vaccine Coordinator Dr Bechara Choucair said of primary care clinicians, “You are the most trusted source of vaccine information for your patients and your communities.” ¹ Surgeon General Vivek Murthy echoed this, saying, “Surveys tell us that 8 in 10 adults turn to their doctors when deciding whether to get a COVID‐19 vaccine.” The Biden administration recognized that leveraging trusted health relationships is important in pandemic response. Primary care is where most people in the United States go for health care and where most have health and health care relationships. ² The quality and duration of those relationships are important to outcomes and are likely important to patient trust, but there are no measures of trust in general use for quality assessment or value‐based payment.

Before examining existing trust measures, it is vital to define the construct. Broadly, trust between people is viewed as an unwritten agreement wherein one person accepts that another person will care for them in a vulnerable situation. ³ , ⁴ This general definition of trust can be found in all facets of health care, from institutional trust in health systems (hospitals, insurers, etc.) to personal trust in caretakers (physicians, nurses, etc.). ⁵ More specifically, trust in a known physician has the added effect of being “based primarily on personal experience and individual personality, than trust in a health plan or trust in doctors in general.” ³ Researchers’ past attempts at measuring how physician's achieve trust focused on five domains: fidelity (caring for the patient and avoiding conflicts of interest), competence, honesty, confidentiality, and global trust (trust that combines elements from different domains). ⁶ Trust requires clinicians to be knowledgeable and competent but also warm and caring, so it has both outcome and relationship components. ⁷ , ⁸ Trust measures may need to address the multiple domains of trust in order to target improvement or to refine understanding of relationship to outcomes.

Achieving trust is associated with increased adherence to chronic care treatments and recommended preventive services in low‐income African American women. ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ Furthermore, trust has been demonstrated to be vital to reducing health inequities, particularly those related to social determinants of health. ² Trust can reduce anxiety and make it possible for patients to discuss mental health issues with their clinician. ⁷ , ¹⁰ , ¹⁵ Trust is also useful in high‐stakes health situations, like the COVID‐19 pandemic, as health care clinicians are often the most trusted advisor for vaccination decisions. ¹⁶ However, a 2014 study found that the United States was tied for 24th place internationally with only 58% of adults in the United States agreeing with the statement, “All things considered, doctors in [your country] can be trusted.” ¹⁵

The US ranking was not specific to primary care but, given the abundance of evidence for trust as outcome and as facilitator of outcomes, reliable measures of trust in primary care could permit assessment and improvement efforts. Having such measures could improve understanding of clinician and practice characteristics associated with increased trust. Threats to and facilitators of trust are also important to understand. For example, whether trust is higher with patient–physician concordance of race/ethnicity/gender/class has mixed results. ¹⁷ , ¹⁸ , ¹⁹ Shifts in health care delivery and presence of technology in health care have also created potential challenges for achieving trust: increased awareness of high‐profile medical errors, shifts to financial incentives for clinicians, and patients using the internet for health concerns that result in them getting answers different than the ones they received from their physician are all factors that may decrease trust in physicians. ¹⁰ , ²⁰ , ²¹ , ²² There are mechanisms for increasing trust. Trust has been associated with better continuity of care, patient satisfaction, and increased patient involvement in decision making. ¹⁰ , ¹¹ , ¹² , ¹³ , ²² , ²³ Longer physician relationships, increased physician choice, and trust in the managed care organization have been shown to be correlated with improved patient's trust in a physician. ²⁴ Furthermore, related to primary care high‐value functions, better care continuity, clinician communication, comprehensiveness, and care coordination of specialty care are all associated with higher trust in primary care clinicians. ²⁵ Having measures that assess trust in a primary care clinician and identify gaps can lead to efforts to improve it, looking for existing threats and in changing how clinicians care for patients.

Most existing clinical trust measures were developed and tested two decades ago. The rapid pace of practice/system consolidation and very high rates of employment of newly trained physicians are changing clinician autonomy and ability to support clinical relationships, continuity, and comprehensiveness, all of which may affect trust. ²⁶ , ²⁷ The capacity to measure trust and to understand its association with practice or clinician characteristics may be important to improving it and improving outcomes.

The Federal Quality Payment Program seeks to pay for value, and relationship to cost and utilization is a priority. Furthermore, it will be important to understand whether or how other measures are associated with trust to further support adoption and implementation of trust measures. For example, the noted relationship between trust and measures of continuity may reinforce the need for both because continuity is associated with good patient outcomes. We sought to assess existing trust measures, particularly trust in a primary care physician, that could be further tested for wide use for assessment, improvement, and value‐based payment.

Following this paper, our intention is to select a trust measure for implementation in our national primary care Qualified Clinical Data Registry for further assessment of clinician and patient perceptions of value, relationships to outcomes, and relationship with other high‐value primary care measures. If a trust measure offers unique and substantial value, our registry has authority from the Centers for Medicare & Medicaid Services to use it for the Quality Payment Program and to nominate it for broader use.

Methods

Literature Search

We conducted a scoping review. We chose to conduct a scoping rather than systematic review due to the lack of prior comprehensive research on trust measures, the exploratory nature of discovering trust measures, and the intention to do post hoc search for evidence of reliability and validity testing. Discovery, rather than refinement, was a primary goal because we know we will need to rigorously test candidate measures. With assistance and recommendations from a professional librarian, we used the PubMed database to search for any relevant papers. Our initial search included “trust” combined with keywords “metrics,” “measurement,” “measurements,” “measuring,” “index,” or “scale” (Appendix Table A1). We limited the search to health measures by using the terms “health care,” “health,” or “medicine.” Medical Subject Headings search was conducted to expand the search to related subheadings. We did not restrict the search to measures that used “trust” in their name or even in the measure construct. This process yielded 2,241 papers. A title and abstract scan was conducted, as well as looking at references, to identify papers that developed or explained trust measures in the medical field. We then read and analyzed available measures, prioritizing measures with a utility to primary care, specifically measures that recognize the continuous, ongoing relationship aspect of primary care in contrast to other specialized forms of care. We further restrained our search by eliminating papers that measured trust in health care apart from clinicians and focused on those that assessed for relationship with outcomes in or affected by primary care. We read for assessments of internal consistency, validity, and reliability as well as relationships to prior measures. Measures that were thoroughly investigated, rather than just produced and used, were prioritized. After narrowing our analysis to four measures that suggested usefulness to primary care, we conducted an additional PubMed search for any instances of those measures being used in primary care, being assessed against outcomes, and being used in English‐speaking settings. This was done by keyword search of the measures’ full name and acronym. The key features of each scale and differences between them were summarized.

Endorsed Clinical Quality Measure Search

We also searched the Centers for Medicare & Medicaid Services Measures Inventory Tool (CMIT) and National Quality Forum's (NQF) measure inventory for trust measures. The CMIT returned hospice and palliative care measures, none of which specifically address trust or consider a patient's perspective on a trusting relationship with the care team. The NQF search returned three measures, none of which are currently endorsed or eligible for value‐based payment programs. Two measures include questions that the developers state are linked to inspiring trust or building trust; and the third measures trust in medical care safety for hospitalized children. None of these measures were returned by our literature search and none met our criteria for assessing trust in primary care clinicians.

Development of Framework/Suggestions

The resulting trust measures were assessed for further development and testing for assessing trust in primary care. Gaps in the previous measures, as well as features of primary care that must be considered when constructing a measure of trust for it, are reported to provide a foundation for future development. After comparing the utility of each measure, we produced a recommendation for the development and use of a future measure in value‐based payment systems. We identified which existing measures have the most potential to serve this purpose effectively.

Results

The described process yielded 59 papers and 57 measures of trust in health care along with 2 systematic reviews. ²⁴ , ²⁶ However, 35 of the papers consisted of measures of trust in something other than physicians, such as hospitals, health insurers, and researchers. Eliminating papers that measured trust in nonphysicians limited our assessment to 22 measures, 9 of which were previously evaluated in the systematic reviews. Five of 13 measures not included in the systematic reviews predated the systematic reviews and may have been excluded, whereas the other 8 were developed after the systematic reviews. After examining the systematic reviews and later measures and scanning each scale for its utility to primary care, two measures focused on primary care physicians and two measures that focused on any clinician in a primary care setting were adequately tested, validated, and used in various studies to allow us to perform further evaluation (Table 1). The four scales are explained in more detail below, including their questions (Table 2), and related studies (Table 3). The process of identifying these papers is outlined in a PRISMA‐style flowchart (Figure 1).

Table 1.

Description of Trust Scales

Name of Scale	Versions a	Author, Year	Target of Measure	Number of Items	Conceptualization of Trust	Hall's Description of Items 6	Number of Studies in Primary Care Setting
Trust subsection of Primary Care Assessment Survey	None	Safran et al., 1998 28	Primary care physician	8 b	Trust measures should assess a primary care physician's integrity, competence, and role as the patient's provider	2 questions on fidelity, 3 honesty, 1 confidentiality, and 2 global trust	22
Trust in Physician Scale	1. Original scale 2. Slight modification of one item	1. Anderson and Dedrick, 1990 8 2. Thom et al., 1999 29	Primary care physicians	11	Interpersonal trust: a person's belief that the physician's words and actions are credible and can be relied on	4 questions on fidelity, 2 honesty, 1 confidentiality, 1 competence, 2 behavior, and 1 global trust	10, including Thom et al., 1999
Health Care Relationship Trust Scale	1. Original scale 2. Revised scale with elimination of two items	1. Bova et al., 2006 10 2. Bova et al., 2012 30	Primary care health professionals	1. 15 2.13	Collaborative trust is unique to health care providers and consists of six key components: knowledge sharing, emotional connection, professional connection, respect, honesty, and partnership	Not given	3, including Bova et al., 2012
Wake Forest Physician Trust Scale	1. Original scale 2. Revised scale	Hall et al., 2002 6	Primary care providers	1.10 2.5	5 domains were fidelity, honesty, competence, confidentiality, and global trust	2 questions on fidelity, 3 competence, 1 honesty, and 4 global trust	4

^a Not including language translation versions.

^b 51 total items, 8 in the trust subscale

Table 2.

Items from Trust Scales

Scale	Items
PCAS 28	1. I can tell my doctor anything 2. My doctor sometimes pretends to know things when he/she is really not sure 3. I completely trust my doctor's judgments about my medical care 4. My doctor cares more about holding costs down than doing what is needed for my health 5. My doctor always tells me the truth about my health, even if there was bad news 6. My doctor cares as much as I do about my health 7. If a mistake was made in my treatment, my doctor would try to hide it from me 8. All things considered; how much do you trust your doctor
TIPS 8 , 29	1. I doubt that my doctor really cares about me as a person 2. My doctor is usually considerate of my needs and puts them first 3. I trust my doctor so much I always try to follow his/her advice 4. If my doctor tells me something is so, then it must be true 5. I sometimes distrust my doctor's opinion and would like a second one 6. I trust my doctor's judgments about my medical care 7. I feel my doctor does not do everything he/she should for my medical care 8. I trust my doctor to put my medical needs above all other considerations when treating my medical problems 9. My doctor is a real expert in taking care of medical problems like mine a 10. I trust my doctor to tell me if a mistake was made about my treatment 11. I sometimes worry that my doctor may not keep the information we discuss totally private
HCR Trust Scale 10 , 30	How often does your HCP: 1. Talk over your head? b 2. Discuss options and choices with you before health care decisions are made? My HCP is: 3. Committed to providing the best care possible 4. Sincerely interested in me as a person 5. An excellent listener My HCP: 6. Accepts me for who I am 7. Tells me the complete truth about my health‐related problems 8. Treats me as an individual 9. Makes me feel that I am worthy of his/her time and effort 10. Takes the time to listen to me during each appointment I feel: 11. That other patients get better care from their HCPs b 12. Comfortable talking to my HCP about my personal issues 13. Better after seeing my HCP How often: 14. Do you think about changing to a new HCP? 15. Does your HCP consider your need for privacy?
WFTS 6 , 11	1. Your doctor will do whatever it takes to get you all the care you need 2. Sometimes your doctor cares more about what is convenient for [him or her] than about your medical needs c 3. Your doctor's medical skills are not as good as they should be 4. Your doctor is extremely thorough and careful c 5. You completely trust your doctor's decisions about which medical treatments are best for you c 6. Your doctor is totally honest in telling you about all of the different treatment options available for your condition c 7. Your doctor only thinks about what is best for you 8. Sometimes your doctor does not pay full attention to what you are trying to tell [him or her] 9. You have no worries about putting your life in your doctor's hands 10. All in all, you have complete trust in your doctor c

Abbreviations: HCP, health care professional; HCR, Health Care Relationship; PCAS, Primary Care Assessment Survey; TIPS, Trust in Physicians Scale; WFTS, Wake Forest Physician Trust Scale.

^a Modified in 1999 version to “My doctor is well qualified to manage (diagnose and treat or make an appropriate referral) medical problems like mine.”

^b Eliminated in 2012 modification.

^c Included in 2005 abbreviated version.

Table 3.

Descriptions of Studies that Used Trust Scales in English‐Speaking Primary Care Settings

Scale	Study	Setting	Sample Size	Key Trust and Outcomes Findings	Scale Properties Examined
PCAS	Blair et al., 2013 31	Primary care patients	2,908	Trust was not related to a physician's implicit or explicit bias	None
	Burge et al., 2011 32	English‐ and French‐speaking patients	645	Same as scale properties	The PCAS trust subscale was highly correlated with the PCAS and other continuity scales, apart from the PCAS visit‐based continuity subscale
	Haggerty et al., 2011 33	English‐ and French‐speaking patients	645	Same as scale properties	Examined bias in measure: urban respondents had higher trust scores than rural respondents, but no difference was observed for trust scores based on level of education
	Haggerty et al., 2011 34	English‐ and French‐speaking patients	645	Same as scale properties	F‐test of discrimination = 96.9; Cronbach alpha internal consistency = 0.88
	Vannoy and Robbins, 2011 35	Adult primary care patients	1,776	Trust was not associated with the likelihood of a patient having suicide‐ or depression‐related discussions with the physician	None
	Garg et al., 2010 36	Mothers of children seeing pediatricians	482	Trust subscale combined with three other PCAS subscales to create a singular score, so no isolated trust conclusions can be drawn	None
	Saitz et al., 2008 37	Primary care patients with alcohol misuse	280	Trust was associated with more alcohol counseling but did not influence the odds of drinking risky amounts	None
	Boyd et al., 2008 38	Older patients of general internists who are chronically ill	150	The implementation of guided care (“an enhancement to primary care that incorporates the operative principles of chronic care innovations”) appeared to have no significant effect on trust	None
	Kim et al., 2007 39	primary care patients with alcohol‐, cocaine‐, and/or heroin‐use disorders	183	Trust was associated with lower alcohol addiction severity and lower odds of substance abuse but not lower drug addiction severity	None
	Duberstein et al., 2007 40	Primary care patients	4,616	None	None
	Shadmi et al., 2006 41	Patients of general internists aged 65 years and older	120	Patients with high morbidity had lower trust than patients with moderate morbidity	None
	Fiscella et al., 2004 42	Primary care patients	4,746	Longer visit lengths were associated with higher trust	Higher scores on the Measure of Patient‐Centered Communication, which measures physician behaviors, were associated with higher trust
	Hartley, 2002 43	Primary care patients	227	A physician's comprehensive knowledge of the patient was not associated with trust	None
	Murphy et al., 2001 44	Primary care patients	4,108	Trust scores declined during the 3‐year study from 1996 to 1999	None
	Murray et al., 2000 45	Primary care patients	6,810	Trust did not appear to change by physician workload between part‐time and full‐time physicians	None
	Safran et al., 2001 46	Primary care patients	3,052	Greater trust was associated with less voluntary disenrollment, or in other words, switching doctors voluntarily	None
	Wilson et al., 2005 47	Medicare patients with at least one prescription drug medication	7,130 (1998), 4,630 (2000)	A better physician–patient relationship, which includes better trust, was associated with less cost‐related skipping of medications. Trust was not reported on independently	None
	Montgomery et al., 2004 48	noninstitutionalized Medicare beneficiaries aged 65 years and older	4,173	The study assessed change in primary care experiences, but trust as assessed by the PCAS did not appear to change during the 2‐year study period from 1998 to 2000	None
	Taira et al., 1997 49	Primary care patients	502	Asian American people rated trust lower than White or African American individuals	None
	Taira et al., 2001 50	Primary care patients	6,092	Asian people had lower scores than White people for trust as well as 8 of the other 10 primary care domains in the study: financial access, organizational access, visit‐based continuity, preventive counseling, comprehensive knowledge of patient, thoroughness of physical examination, communication, and interpersonal treatment. Trust for Latino, African American, and White people did not differ significantly	None
	Safran et al., 1998 51	Primary care patients	7,204	Greater trust was strongly correlated with adherence to medical advice	Of all the PCAS subscales, trust was most strongly associated with satisfaction: trust accounted for 35% of the variance in satisfaction scores
TIPS	Thom et al., 1999 29	Primary care patients	414	Trust was higher for longer patient–physician relationships, patients who actively chose their physician, patients who preferred more physician involvement, and patients who expected their physician to care for a larger proportion of their problems	Creation of modified scale: internal consistency Cronbach alpha = 0.99 and had a good 1‐month test–retest reliability
	Pirotta et al., 2014 52	Primary care patients	7,432	Use of the herbal medicine St. John's wort for mental disorders had no significant association with trust in a patient's general practitioner	None
	Thom et al., 2014 53	Low‐income primary care patients with diabetes, hypertension, and/or hyperlipidemia	441	Having a health coach was shown to increase a patient's trust in their PCP	None
	Nelms et al., 2014 54	Primary care patients who are Medicaid‐enrolled smokers	214	Trust was not associated with any physician‐related behaviors, but higher self‐reported health status was associated with higher trust	None
	Reid et al., 2013 55	Primary care patients aged 14–24 years who used the mental health “mobletype” program	114	The program had no significant influence on trust	None
	Hwong, et al., 2018 20	Primary care patients in Massachusetts	278	Disclosure of industry payments to physicians affected perceptions of individual physician honesty and fidelity, but not perceptions of competence	None
	Martin et al., 201356 a	Primary care patients	227	White patients had higher trust than Black patients, and higher trust was associated with longer visit length and positive physician affect (good communication behaviors)	None
	Moseley et al., 200613 b	Parents of children receiving pediatric care	485	Lower trust scores were associated with being a father, older parent age, parents with private insurance status, parent education greater than high school, not having a child less than 3 years of age, and parents who were not White	Internal consistency alpha = 0.84
	Benkert et al., 200957 c	African American patients of nurse‐practitioner primary care centers	100	Higher trust was associated with greater satisfaction	Internal consistency of 0.85‐0.90 with African American samples
	Clancy et al., 2003 58	Patients with type 2 diabetes	120	Patients in group visits had higher trust compared with patients with usual care	None
	Clancy et al., 2007 59	Patients with type 2 diabetes: a follow‐up to above study	186	Contradictory finding to above study: trust showed no association with group visits as compared to usual care	None
HCR Trust Scale	Bova et al., 2012 30	Primary care patients	431	None	Resulted in the creation of a modified scale: good scores for structural validity and internal consistency, fair score for hypotheses testing
	Mancuso, 2010 60	Primary care patients with diabetes	102	As patient trust increased, glycosylated hemoglobin concentration (HbA1c) decreased, indicating better glycemic control	None
	Kamimura et al., 2020 61	Uninsured primary care patients	489	Trust not assessed independently but rather grouped together for conclusions about the patient–provider relationship	None
WFTS	Coats et al., 2018 62	Primary and specialty care patients with chronic, life‐limiting illnesses	537	Trust was associated with racial/ethnic minority status, lower education, and higher ratings of clinician quality of communication, but not with religion nor income	None, used abbreviated scale
	Dugan et al., 2005 11	Primary care patients (2 samples)	1,064 and 1,045, respectively	Trust correlated with satisfaction with the physician, would recommend to friends and family, general satisfaction with care, no desire to switch to another doctor, number of years under physician's care, number of visits to physician, having enough choice in the selection of the physician, not having had a dispute with the physician, and not having sought a second opinion because of concerns about care. Trust decreased with worse mental and physical health and was not associated with education level or income	Creation of abbreviated scale
	Gupta et al., 2014 63	Primary care patients aged 50–79 years	701	Trust in primary care provider associated with a greater likelihood of colorectal cancer screening	None, used abbreviated scale
	Hall et al., 2002 6	Random sample of insured patients who had been to the doctor twice in the past year	National sample: 959 Regional sample: 1,199	Both samples showed significant correlation to the Kao physician trust scale, insurer trust, general satisfaction with health care, satisfaction with physician, length of time with physician, number of prior visits to doctor, willingness to recommend to friends, and intent to switch physician. Lower physician trust was associated with not enough choice in selecting a physician, prior dispute with physician, having sought a second opinion, and membership in managed care	The regional sample had a 2‐month test‐retest reliability of r = 0.75, scale distribution skewed left, and a positive kurtosis. Similar mean and distribution as TIPS and PCAS
	Kanter et al., 2019 21	Nationally representative sample, with oversampling in Massachusetts and Minnesota	3542, wave one 2711, wave two	Nationwide public disclosure of industry payments may be associated with decreased trust in physicians and in the medical profession	None, used abbreviated scale
	Nguyen et al., 2020 64	Convenience sample grocery store shoppers	survey with 600 racially and ethnically diverse adults in Chicago	Interpersonal trust did not differ by racial or ethnic group	None, used full scale
	Stults et al., 2020 65	Convenience sample of men who have sex with men	National sample: 4,239 men	Of men who have sex with men with a PCP, characteristics associated with lower patient–provider trust included younger age, Asian, bisexual, HIV‐negative‐not‐on‐preexposure prophylaxis, HIV‐unknown, and lower medical literacy	None, used abbreviated scale

Abbreviations: HCR, Health Care Relationship; PCAS, Primary Care Assessment Survey; PCP, primary care physician; TIPS, Trust in Physicians Scale; WFTS, Wake Forest Physician Trust Scale.

^a Only used 5 out of 11 items from TIPS.

^b Used Pediatric Trust in Physician Scale, a revised scale from TIPS; changed wording from “my doctor” to “my child's doctor” and other similar wording changes.

^c Changed questions from “physician” to “provider.”

Figure 1.

PRISMA Diagram of Process of Identifying Trust Measure

Health Care Relationship Trust Scale

The Health Care Relationship (HCR) Trust Scale, created by Bova and colleagues in 2006, is a 15‐item scale that aims to measure trust in primary health care professionals (HCPs) such as physicians, nurse practitioners, and physician assistants. ⁹ Focus groups and experts informed creation of this scale aiming to measure trust in vulnerable health communities and tested later with adults with HIV. The measure introduced the concept of collaborative trust, which the creators say is unique to HCPs and built over time. Collaborative trust consists of six key components: knowledge sharing, emotional connection, professional connection, respect, honesty, and partnership. Validation testing found that the 15‐item scale with five subsections had a Cronbach alpha internal consistency of 0.92 (measure of how well items are related, best scores are above 0.80) and test–retest reliability of 0.59 (on repeat measurement, it does produce similar scores; also, ideally above 0.80). The scale was retested with primary care patients and revised by the original author. ³⁰ The primary care study revealed two items with low factor loadings, and these were eliminated. The resulting 13‐item HCR Trust Scale revised had a Cronbach alpha of 0.96. White patients had higher trust scores than Hispanic patients, and physicians had higher trust scores than medical residents. Trust was weakly correlated with older patient age and a longer length of relationship with the HCP. The Müller systematic review reported that the revised scale had good scores for structural validity and internal consistency and a fair score for hypotheses testing. ⁶⁶ Only two additional studies of HCR Trust Scales in an English‐speaking primary care setting were found (Table 3). ⁶⁰ , ⁶¹

The Primary Care Assessment Survey

The Primary Care Assessment Survey (PCAS), created by Safran and colleagues in 1994, is a 51‐item scale that covers seven different primary care domains, including trust, and was developed specifically for primary care. ²⁸ The trust domain has eight items and attempts to assess a primary care physician's integrity, competence, and role as the patient's provider. The items have been classified as having two questions about fidelity, three about honesty, one about confidentiality, and two global trust. ⁶ The scale trust subsection was found to have a Cronbach alpha of 0.86, equal item variance, and equal item‐scale correlations, meaning that each item contributed an even amount of information about trust and had high correlation with the full PCAS. As for correlation with the other PCAS subscales, the authors found that the trust section's Cronbach alpha exceeded any correlation, indicating the uniqueness of the trust concept. In testing the PCAS global trust question, “All things considered, how much do you trust your doctor,” Hall found that people do not appear to distinguish trust among the dimensions of fidelity, competence, and honesty, that is to say, this global item had the highest correlation to the overall scale (0.80/0.82). ⁶ Across all the PCAS questions, the highest correlation for the trust subscale was with the communication and interpersonal treatment subscales. The trust subscale was significantly correlated with continuity but not highly correlated (correlation of 0.30 for visit‐based continuity and 0.22 for longitudinal continuity), suggesting that continuity may not be a sufficient proxy for trust. Twenty‐one different studies have used the PCAS in a primary care, English‐speaking setting (Table 3).

Trust in Physicians Scale

The Trust in Physicians scale (TIPS) is an 11‐item scale that was developed to measure trust in primary care physicians and authored by Anderson and Dedrick in 1990. ⁸ The scale attempts to assess interpersonal trust, which it defines as “a person's belief that the physician's words and actions are credible and can be relied upon.” The items are categorized as having four questions about fidelity, two about honesty, one about confidentiality, one about competence, two about behavior, and one global. ⁶ Two validation studies produced Cronbach alphas of 0.90 and 0.85 and a test–retest reliability of 0.77. In the first study, they found (older) patient age and less education to be associated with greater trust; race did not modify association. ⁸ In the second study, done in 1999, trust accounted for 38.4% of the variance in satisfaction. ²⁹ The Müller systematic review analyzed the 1999 modification and found the scale to have a fair score for reliability, but poor for internal consistency and hypotheses testing. ⁶⁶ The 1999 study found good one‐month test–retest reliability and an internal consistency of 0.89, and that trust was not associated with education level or gender, but increased with older age. ²⁹ Trust also correlated with satisfaction and perceived humaneness of physician behavior. The TIPS has been used in a primary care, English‐speaking setting in 11 different studies (Table 3).

Wake Forest Physician Trust Scale

The Wake Forest Trust Scale (WFTS), created by Hall and colleagues in 2002, is a 10‐item, Likert‐style scale designed to assess trust in primary care clinicians. ⁶ The WFTS aimed to improve on limitations of previous scales, which its authors defined as conceptual (PCAS) and inconsistencies between the scales (PCAS and TIPS). The measure considered five domains of trust: fidelity, competence, honesty, confidentiality, and global trust. The authors differentiated trust from satisfaction in that satisfaction is experience‐based whereas trust is future‐oriented. The authors omitted the word “trust” from the scale to measure trust based on its independent factors. Items were generated by modifying questions from previous trust scales as well as creating new questions. Following a multistep process including expert review panel, focus groups, pilot testing, and comprehensive statistical and factor analysis, 78 items were reduced to 10. Two questions asked about fidelity, three about competence, one about honesty, and four about global trust. One item came from the TIPS and two from the PCAS; confidentiality was one of the eliminated elements. A study with 959 patients in a national sample and an additional 1,199 patients in a regional sample assessed scale validity and correlation with other factors. ⁶ The scale had a Cronbach alpha of 0.93 (national) and 0.92 (regional), as well as a test–retest reliability of 0.75. Furthermore, the scale was skewed left and had a positive kurtosis (higher clustering at lower scores, more discriminating at higher scores). They also found higher trust to be associated with higher insurer trust, general satisfaction with health care, satisfaction with physician, greater visit continuity, greater willingness to recommend to friends, and lower intention to switch physician. The greater length of time with physician was only significantly associated with trust in the regional sample. Lower physician trust was found to be associated with reduced choice in selecting a physician, prior dispute with physician, having sought a second opinion, and membership in managed care. The scale had a similar mean and distribution as TIPS and PCAS. The authors assert that “the improvements made in this scale make it especially suited to measuring trust in settings where it is likely to be high and where it is important to make subtle distinctions between trust and other related constructs.” ⁶

An abbreviated, 5‐item version of the scale was created in 2005 by the scale's original author to reduce respondent burden and research cost. ¹¹ The abbreviated scale also used national and regional samples to develop and test the psychometric properties of the scale. The new scale had a Cronbach alpha of 0.87 and 0.86 in the two samples and a test–retest reliability of 0.71. Although internal consistency was lower than the original scale, it remained correlated with satisfaction with the physician, would recommend to friends and family, general satisfaction with care, no desire to switch to another doctor, more years under the physician's care, more visits to the physician, having enough choice in the selection of the physician, not having had a dispute with the physician, and not having sought a second opinion. Trust decreased with poorer physical and mental health. The WFTS have been used in seven studies in a primary care, English‐speaking setting (Table 3).

Discussion

Our literature review was purposeful and scoping with the aim of understanding if there are primary care clinician trust measures with sufficient evidence of reliability (high internal consistency and test–retest reliability) and discriminatory capacity (does it offer meaningful trust differences between clinicians) for general use. Ideally such a measure would also be tested for its relationship to other primary care functions, such as continuity and comprehensiveness, as these are becoming standardized measures, as well. Concurrent to our work, the Building Trust Initiative of the American Board of Internal Medicine Foundation, a leading contributor for primary care standards, seeks to examine “issues of trust in health care and how trust contributes to better health care outcomes, increased patient satisfaction and greater physician well‐being, among other cases.” ⁴ The Foundation's efforts focus on how health care systems can build trust rather than on measuring trust in clinicians, but a signal that trust in health care is growing as an important focus. Safran has shown that trust “encourages use of services, facilitates disclosure of important medical information and has an indirect influence on health outcomes through patient satisfaction, adherence, and continuity.” ⁵¹ We share the enthusiasm for the promise of trust as an important measure and policy focus, as a driver of support for relationships in primary care, and as a focus for further research. Despite important leaps in research and development of useful primary care trust measures, none have progressed to general use or endorsement for use in value‐based payment programs. Trust and building trust in physician relationships has been repeatedly mentioned as being important to the COVID‐19 pandemic, but we have no consensus on which measure to assess or improve on. ¹ , ⁶⁷ , ⁶⁸ , ⁶⁹ In the introduction, we summarize evidence for threats and supports of trust in primary care clinicians. As a value‐based payment measure, trust would serve to reinforce clinical and other policies that improve trust and reduce those that threaten trust.

We believe that the TIPS and WFTS scales are the best candidates to address our long‐term intention to implement a trust measure into our national primary care Qualified Clinical Data Registry. The TIPS scale has not been widely tested or used but it uses items derived from PCAS, which was thoroughly tested. The TIPS scale asks about the physician's capacity to care for patients, including items such as putting their health above all else and being truthful. The TIPS also includes competency questions that assess patients’ beliefs in their physician's ability and knowledge but is unclear whether this addition is valuable for assessing trust. The WFTS contains some similar items as PCAS and, like TIPS, includes physician competence as an aspect of trust. The extensive statistical evaluation of the WFTS and its revision and specific development and testing in primary care suggest that the scale may prove to be the most fit for purpose. The scale warrants further testing, empirically and qualitatively, and our national registry offers a good laboratory for both. The HCR Trust Scale is of interest because it is not exclusive to physicians, but we found it to be insufficiently tested.

Our scoping review may have missed other measures of trust in primary care clinicians, particularly those that were not explicit about testing in primary care or not in English‐speaking countries or settings. Although the methods used did cast a broad net and assess for testing of the measures for internal consistency, validity, reliability, relationships to prior measures and to outcomes, it was not a systematic review and did not aim to produce a definitive best‐in‐class measure of trust. Despite these limitations, we do believe that two measures are worthy of new testing for general use. Hall's finding that global trust assessment questions perform nearly as well as more nuanced scales with multiple components suggests that a simpler solution may be available, particularly if assessed in a broader primary care measure effort.

Conclusion

The renewed recognition of the value of trusted, healing relationships in primary care and national interest in increasing trust in health care should renew efforts to find an effective way to measure it. Implementing a trust measure in primary care practices can improve our assessment of the physician–patient relationship. It also has implications for reducing health inequities for marginalized groups. Based on this review we recommend that the revised versions of TIPS and WFTS should be retested in conjunction with other high‐value primary care measures, including patient‐reported outcome measures. This would include psychometric testing for correlations and unique contributions of elements. More work must be done, but the prior development of trust measures provides an invaluable set of resources on which to build. Incorporating a future measure of trust into a wider array of primary care measures will also be important to improve its feasibility and utility. Research that assesses what influences trust, as well as what health outcomes are associated with trust, should be prioritized.

Funding/Support: None declared

Conflict of Interest Disclosures: None declared.

   

Table A1.

Search Terms

Topic	Search terms used in PubMed (combined with AND)
Trust	(trust[tiab] OR confidence OR faith)
Measurements	(metrics[tiab] OR measurement[tiab] OR measurements[tiab] OR measuring[tiab])
Healthcare	(primary health care[mesh] OR primary care[tiab] OR family practice OR family medicine[tiab])
Date range searched	PubMed was searched through February 2021
Results	2241 results