Abstract
Objective
To describe the development and psychometric testing of the Multidimensional Trust in Health Care Systems Scale (MTHCSS).
Methods
Scale development occurred in 2 phases. In phase 1, a pilot instrument with 70 items was generated from the review of the trust literature, focus groups, and expert opinion. The 70 items were pilot tested in a sample of 256 students. Exploratory factor analysis was used to derive an orthogonal set of correlated factors. In phase 2, the final scale was administered to 301 primary care patients to assess reliability and validity. Phase 2 participants also completed validated measures of patient-centered care, health locus of control, medication nonadherence, social support, and patient satisfaction.
Results
In phase 1, a 17-item scale (MTHCSS) was developed with 10 items measuring trust in health care providers, 4 items measuring trust in health care payers, and 3 items measuring trust in health care institutions. In phase 2, the 17-item MTHCSS had a mean score of 63.0 (SD 8.8); the provider subscale had a mean of 40.0 (SD 6.2); the payers subscale had a mean of 12.8 (SD 3.0); and the institutions subscale had a mean of 10.3 (SD 2.1). Cronbach’s α for the MTHCSS was 0.89 and 0.92, 0.74, and 0.64 for the 3 subscales. The MTHCSS was significantly correlated with patient-centered care (r = .22 to .62), locus of control—chance (r = .42), medication nonadherence (r = −.22), social support (r = .25), and patient satisfaction (r = .67).
Conclusions
The MTHCSS is a valid and reliable instrument for measuring the 3 objects of trust in health care and is correlated with patient-level health outcomes.
Key Words: Multidimensional Trust in Health Care Systems Scale, interpersonal trust, health care, psychometrics, trust, health services research
BACKGROUND
Trust is operationally defined as the optimistic acceptance of a vulnerable situation in which the trustee believes the trustor will take care of the trustee’s interests1. There are 2 distinct aspects of trust. Interpersonal trust is the type of trust that develops when there is opportunity to repeatedly test how dependable an individual is over time. In contrast, social trust refers to trust in a collective institution and is typically influenced by past experiences, the media, and general social confidence in the particular institution2.
Trust in health care providers, health care institutions, and health care payers appears to be influenced in varying degrees by interpersonal trust and social trust. Trust also has both behavioral and attitudinal components. An example of the attitudinal component of trust is the willingness of a patient to forgive medical errors if there is the perception that the provider, institution, or payer was caring and meant well. Similarly, patients may have high levels of trust in a provider, an institution, or a health payer even though the care provided is objectively inferior1. Distrust conversely is not necessarily the absence of trust. Distrust may be defined as having anxious or pessimistic views of motivation and expected results3 or a substitute or complement to trust such that an initially distrustful patient may become very trusting if initial experiences are positive4–5.
The object of trust in the health care setting is who or what the patient trusts and what about them is trusted1. There are 3 main objects of trust in the health care setting: (1) health care providers (physicians, nurse practitioners, physician assistants), (2) health care institutions (hospitals and clinics), and (3) health care payers (managed care plans, private insurers, government insurance). Trust may differ depending on the object of trust and the level of trust of 1 object may be influenced by levels of trust of other objects. For instance, the trust a patient has in a particular health care provider may influence the level of trust the patient has on the health care provider’s institution. Similarly, patients may be trusting of a particular provider but distrustful of the health care payer.
Numerous studies have shown that there are multiple dimensions of trust1,6–10. Most studies suggest that at least 4 dimensions are important. These include agency/fidelity, competence, honesty, and confidentiality1. Agency/fidelity is the pursuit of a patient’s best interest and consists of caring, respect, advocacy, and avoiding conflicts of interest. Competence is the avoidance of cognitive or technical mistakes while producing the best results. Honesty demands telling the truth and avoiding intentional falsehoods such as outright lies, half-truths, or deception by silence. Finally, confidentiality involves protection and proper use of sensitive information. It appears that these dimensions, although conceptually different, are so correlated that they perform as a single construct1.
Although many scales are available to measure trust in specific components of the health care system,6–19 few currently available scales are designed to measure trust in health care providers, trust in health care institutions, and trust in health payers concurrently. Therefore, the goals of this study were threefold: (1) to develop a Multidimensional Trust in Health Care Systems Scale (MTHCSS) that captures the 3 major objects of trust including trust in health care providers, trust in health care institutions, and trust in health payers; (2) to examine the scale’s psychometric properties; and (3) to assess correlations between the multidimensional scale of trust and patient-level health outcomes, including patient-centered care, locus of control, medication nonadherence, social support, and patient satisfaction.
METHODS
The study was conducted in 2 phases—phase 1 (scale development) and phase 2 (evaluation of the final instrument).
Phase 1: Scale Development
Conceptual Model We developed a conceptual model after a review of the literature related to objects of trust/distrust in the health care system. In our conceptual model, we defined trust in health care systems as comprised of trust in health care providers, health care insurers, and health care institutions. Patient characteristics included in the model were age, gender, marital status, educational level, insurance status, income, and having a usual source of care. Consistent with the findings of prior studies, we also identified race/ethnicity as an important variable that influences the level of trust patients have in health care systems13,20–27.
Focus Groups A convenience sample of 48 subjects participated in 9 focus groups segmented by race and gender. Participants were adults recruited from a primary care clinic of an academic medical center. Participants received a $25 honorarium for sessions that lasted approximately 2 hours. The mean age of the participants was 46 years. About 81% were African American, and 42% were women. Two experienced qualitative researchers were compensated to serve as focus group moderators. One moderator was white, whereas the other was African American. Moderators were matched to the race/ethnicity of the participants. The primary author (Egede) attended most of the sessions. Each moderator used a written discussion guide.Open-ended questions were used to initiate conversation before specific probes were introduced. The moderators’ transcripts were reviewed after each session and used to modify the conduct of subsequent sessions. Sessions were conducted until “theoretical saturation” was reached. Full and accurate transcriptions of audiotapes, moderator, and research assistant’s notes were made. The transcripts were read by the moderators and the primary author (Egede) to identify themes related to trust or distrust of the medical care system. The primary author (Egede) made the final decision about primary themes and items to include in the pilot scale.
Item Generation and Selection Following the focus groups, items were generated and selected for pilot testing. We developed a pilot trust scale that included items from previously published trust scales6–19 and items created and added based on information collected during the focus groups. Some items selected from previously published scales were modified from their original form to make them pertinent to the 3 major objects of trust (providers, payers, and institutions). The initial items were reviewed by 2 experts familiar with the current literature on trust to determine the relevance of the items for measuring the construct of trust. After expert review, 70 items consisting of statements related to trust in health care providers, health care payers, and health care institutions were selected for pilot testing. Items were scored on a 5-point Likert scale with scores ranging from 5 (strongly agree) to 1 (strongly disagree). A summary score consisting of the sum of individual items was created, so that higher total and subscale scores represented greater trust in health care systems.
Pilot Scale Testing The 70-item pilot trust scale was administered to a convenience sample of students at a university in the southeastern United States via a web-based survey (pilot sample). All enrolled undergraduate students were e-mailed an invitation to participate in the study. The e-mail contained a link to the web-based survey. Approximately 2,000 students were contacted. Approximately 1,000 e-mails were returned as invalid addresses, thus ~1,000 students received the initial e-mail. These students with valid e-mail addresses were solicited via 2 additional e-mails sent at 1-week intervals. Of this number, 257 completed the web-based survey, resulting in a response rate of approximately 26%. All responses were collated and stored in Statistical Package for the Social Sciences (SPSS 14.0)28.
Data reduction techniques were employed to construct a parsimonious multidimensional trust scale, which combines the strengths of other trust scales and incorporates the 3 objects of trust. An exploratory factor analysis (principal component analysis with Varimax Rotation) was used to derive a set of correlated factors that explained the maximum amount of variation among all the scale items. We used the Kaiser–Guttman criterion (eigenvalues greater than 1.0) to decide the number of factors and associated items that would be retained29. Item–item and item–total scale correlations analysis were completed to ensure that internal consistency and reliability existed in the scale. Item selectivity correlations were calculated for each item on the pilot scale as a measure of item–item correlation. We checked individual item loadings with all identified factors. Items were retained if they only loaded on 1 factor with a factor loading of at least 0.40. In addition, items were selected if they correlated at greater than or equal to .5 with at least 1 factor and did not correlate at greater than .3 with the other factors29. Cronbach’s alphas were calculated for each subscale and the total scale as a measure of internal consistency and reliability (item–scale correlation) or the extent to which the items measure the same aspects of the same characteristic.
Phase 2: Evaluation of Final Instrument
In phase 2, we administered the revised final scale to a convenience sample of 301 consecutive patients (final sample) attending a primary care clinic at an academic medical center in the southeastern United States. Patients were recruited over a 10-week period. The final scale was based on the constructs (factors) and items that emerged from phase 1. The same Likert scale scoring was used in the final scale as in the pilot scale. A confirmatory factor analysis (principal component analysis with Varimax Rotation) was performed to derive a set of correlated factors that explain the maximum amount of variation among all the scale items. We used the same procedures for factor retention and item analysis to examine internal consistency and reliability. Item–item and item–total scale correlations analysis were again completed to ensure that internal consistency existed in the scale. Mean trust scale summary scores and mean subscale scores were compared by sample demographic characteristics (age, sex, race/ethnicity, marital status, educational level, insurance status, and income).
Validated measures of patient-centered care, locus of control, medication nonadherence, social support, and patient satisfaction were completed to examine criterion related validity of the scale. Patient-centered care was measured with the Modified Picker Survey,30–32 health locus of control was measured with the Multidimensional Health Locus of Control Scale,33 medication nonadherence was measured with the Morisky Scale,34 social support was measured with the Medical Outcomes Study (MOS) Social Support Survey,35 and patient satisfaction was measured with a previously validated general scale to measure satisfaction/dissatisfaction with health care. We selected these variables because each has been associated with trust14. We hypothesized that these variables would be associated with levels of trust.
The studies were reviewed and approved by our Institutional Review Board for Human Research.
RESULTS
Phase 1: Scale Development
A convenience sample of 257 students responded to the online survey. The response rate was 26% of the students invited to participate. The demographic characteristics of the sample are reported in Tables 1 and 2. Briefly, 81% were white, 61% were greater than 25 years old, 75% were female, and 89.1% saw a physician in the past 12 months in the pilot sample (see Table 1).
Table 1.
Pilot Sample Characteristics
| Characteristics | Percent |
|---|---|
| Pilot sample (n = 257) | |
| Age | |
| <25 years | 38.7 |
| 25–29 years | 32.8 |
| 30–39 years | 19.1 |
| 40+ years | 9.4 |
| Sex: women | 75.0 |
| Race/ethnicity | |
| White | 81.2 |
| African American | 9.0 |
| Hispanic/other | 9.8 |
| Marital status | |
| Single | 59.2 |
| Married | 34.5 |
| Separated/divorced | 6.3 |
| Educational level | |
| Freshman | 19.1 |
| Sophomore | 24.6 |
| Junior | 20.1 |
| Senior | 36.2 |
| Insurance status | |
| Individual plan | 35.5 |
| Group plan | 61.3 |
| Uninsured | 3.1 |
| Income | |
| <$5,000 | 20.1 |
| $5,000–$19,999 | 17.7 |
| $20,000–$34,999 | 26.1 |
| $35,000+ | 36.1 |
| Usual source of care: Yes | 63.8 |
| Saw health provider past 12 months: Yes | 89.1 |
| Used Emergency room past 12 months: Yes | 11.1 |
Table 2.
Final Sample Characteristics
| Characteristics | Percent |
|---|---|
| Final sample (n = 301) | |
| Age | |
| <35 years | 10.9 |
| 35–49 years | 29.0 |
| 50–64 years | 39.6 |
| 65+ years | 20.5 |
| Sex: women | 64.5 |
| Race/ethnicity | |
| White | 48.7 |
| African American | 48.3 |
| Hispanic/other | 3.0 |
| Marital status | |
| Single | 16.8 |
| Married | 54.4 |
| Separated/divorced | 18.5 |
| Widowed | 10.4 |
| Educational level | |
| <High school graduate | 15.4 |
| High school graduate | 29.0 |
| <College graduate | 22.5 |
| ≥College graduate | 33.1 |
| Insurance status | |
| Private insurance | 60.3 |
| Government insurance | 33.2 |
| Uninsured | 6.4 |
| Income | |
| <$10,000 | 23.3 |
| <$25,000 | 12.5 |
| <$35,000 | 10.8 |
| <$50,000 | 13.3 |
| <$75,000 | 20.4 |
| <$100,000 | 11.5 |
| $100,000+ | 8.2 |
| Usual source of care: Yes | 86.9 |
| Saw health provider past 12 months: Yes | 94.6 |
| Used Emergency room past 12 months: Yes | 11.1 |
| Number of health provider visits in past 12 months | |
| 0/1 | 7.4 |
| 2 | 17.9 |
| 3 | 13.6 |
| 4 | 17.5 |
| 5+ | 43.6 |
| Number of emergency department visits in past 12 months | |
| 0 | 7.9 |
| 1 | 55.6 |
| 2 | 11.9 |
| 3 | 9.5 |
| 4+ | 15.1 |
Internal Consistency and Reliability The exploratory factor analysis (principal component analysis with Varimax Rotation) identified a 17-item MTHCSS with 3 factors (subscales). The eigenvalues for the 3 components were 6.29, 2.40, and 1.30, respectively. These 3 components of the scale explained 59% of the variance. The loadings for the rotated items ranged between 0.656 and 0.834 (See Table 3). All individual items correlated at greater than or equal to .5 with at least 1 factor and did not correlate at greater than .3 with the other 2 factors29. The factors (subscales) for the revised scale were: trust in the health care providers subscale (10 items), trust in the health care payers subscale (4 items), and trust in health care institutions subscale (3 items). Cronbach’s α for the total scale was 0.86 and 0.90, 0.75, and 0.72, respectively, for the 3 subscales. The Cronbach’s α for the total and 3 subscales were within the generally recommended range of 0.70–0.9029. The psychometric properties of each individual item and the total scale of the MTHCSS in the pilot sample are reported in Table 4.
Table 3.
Component Loadings of the Principal Component Analysis after Varimax Rotation for Pilot (P) and Final Sample (F)
| Items | Components | ||||||
|---|---|---|---|---|---|---|---|
| 1 | 2 | 3 | |||||
| P | F | P | F | P | F | ||
| Subscale 1: Trust in Health Care Providers | |||||||
| 1 | My health care provider is usually considerate of my needs and puts them first | 0.73 | 0.75 | 0.02 | −0.18 | 0.21 | 0.07 |
| 2 | I have so much trust in my health care provider that I always try to follow his/her advice | 0.74 | 0.80 | −0.08 | −0.19 | 0.19 | 0.01 |
| 3 | I trust my health care provider so much that whatever he/she tells me, it must be true | 0.67 | 0.75 | 0.10 | −0.11 | 0.04 | −0.03 |
| 4* | Sometimes, I do not trust my health care provider’s opinion and therefore I feel I need a second one | 0.67 | 0.48 | 0.18 | −0.10 | −0.18 | 0.13 |
| 5 | I can trust my health care providers judgments concerning my medical care | 0.80 | 0.77 | −0.01 | −0.18 | 0.13 | −0.08 |
| 6 | My health care provider will do whatever it takes to give me the medical care that I need | 0.68 | 0.83 | 0.14 | −0.19 | 0.21 | 0.001 |
| 7 | Because my health care provider is an expert, he is able to treat medical problems like mine | 0.66 | 0.77 | −0.07 | −0.23 | 0.19 | −0.17 |
| 8 | I can trust my health care provider’s decisions on which medical treatments are best for me | 0.79 | 0.83 | 0.05 | −0.22 | 0.18 | −0.13 |
| 9 | My health care provider offers me the highest quality in medical care | 0.77 | 0.83 | 0.10 | −0.20 | 0.26 | 0.05 |
| 10 | All things considered, I completely trust my health care provider | 0.83 | 0.86 | 0.05 | −0.18 | 0.12 | −0.03 |
| Subscale 2: Trust in Health Care Payers | |||||||
| 11 | Health care payers are good at what they do | 0.13 | 0.39 | 0.74 | 0.53 | 0.09 | −0.22 |
| 12 | When needed, health care payers will pay for you to see any specialist | 0.01 | 0.38 | 0.66 | 0.59 | 0.20 | −0.21 |
| 13 | When questioned about what treatments are covered, health care payers are honest with their answers | 0.03 | 0.39 | 0.82 | 0.65 | 0.03 | −0.27 |
| 14 | Health care payers will pay for everything they are supposed to, including treatment that is expensive | 0.02 | 0.32 | 0.74 | 0.60 | 0.12 | −0.32 |
| Subscale 3: Trust in Health Care Institutions | |||||||
| 15* | Health care institutions only care about keeping medical costs down, and not what is needed for my health | 0.12 | 0.25 | 0.12 | 0.29 | 0.74 | 0.67 |
| 16 | Healthcare institutions provide the highest quality in medical care | 0.23 | 0.49 | 0.15 | 0.41 | 0.77 | 0.42 |
| 17 | When treating my medical problems, health care institutions put my medical needs above all other considerations, including costs | 0.27 | 0.48 | ,22 | 0.51 | 0.72 | 0.36 |
*Items 4 and 15 are reverse scored
Table 4.
Items and Statistical Measures of the MTHCSS in Pilot Sample
| Item number | Wording of items | Means (SD) | Item selectivity | Cronbach’s a without item |
|---|---|---|---|---|
| Subscale 1: Trust in Health Care Providers (α = 0.90) | 26.0 (6.0) | |||
| 1 | My health care provider is usually considerate of my needs and puts them first | 2.2 (0.7) | 0.58 | 0.74 |
| 2 | I have so much trust in my health care provider that I always try to follow his/her advice | 2.5 (1.0) | 0.52 | 0.74 |
| 3 | I trust my health care provider so much that whatever he/she tells me, it must be true | 3.5 (0.9) | 0.49 | 0.74 |
| 4* | Sometimes, I do not trust my health care provider’s opinion and therefore I feel I need a second one | 3.4 (0.9) | 0.48 | 0.81 |
| 5 | I can trust my health care providers judgments concerning my medical care | 2.3 (0.7) | 0.60 | 0.74 |
| 6 | My health care provider will do whatever it takes to give me the medical care that I need | 2.5 (0.9) | 0.60 | 0.73 |
| 7 | Because my health care provider is an expert, he is able to treat medical problems like mine | 2.4 (0.8) | 0.48 | 0.74 |
| 8 | I can trust my health care provider’s decisions on which medical treatments are best for me | 2.3 (0.6) | 0.64 | 0.73 |
| 9 | My health care provider offers me the highest quality in medical care | 2.5 (0.8) | 0.67 | 0.73 |
| 10 | All things considered, I completely trust my health care provider | 2.4 (0.8) | 0.63 | 0.73 |
| Subscale 2: Trust in Health Care Payers (α = 0.75) | 13.3 (3.1) | |||
| 11 | Health care payers are good at what they do | 3.2 (1.0) | 0.35 | 0.75 |
| 12 | When needed, health care payers will pay for you to see any specialist | 3.4 (1.1) | 0.28 | 0.76 |
| 13 | When questioned about what treatments are covered, health care payers are honest with their answers | 3.1 (1.0) | 0.33 | 0.75 |
| 14 | Health care payers will pay for everything they are supposed to, including treatment that is expensive | 3.6 (1.0) | 0.28 | 0.76 |
| Subscale 3: Trust in Health Care Institutions (α = 0.72) | 8.4 (2.3) | |||
| 15* | Health care institutions only care about keeping medical costs down, and not what is needed for my health | 2.6 (1.0) | 0.39 | 0.81 |
| 16 | Healthcare institutions provide the highest quality in medical care | 2.6 (0.8) | 0.46 | 0.74 |
| 17 | When treating my medical problems, health care institutions put my medical needs above all other considerations, including costs | 3.2 (1.0) | 0.49 | 0.74 |
| Total scale (α = 0.86) | 47.6 (8.5) |
*Items 4 and 15 are reverse scored
Phase 2: Evaluation of Final Instrument
Three hundred one patients participated in the phase 2 study (final sample). The demographic characteristics of the final sample are reported in Table 2. Briefly, 49% were white, and 48% were African American. Approximately 65% were female, 60% were at least 50 years old, and 95% saw their provider in the past 12 months. The mean MTHCSS score for the final sample was 63.0 (SD 8.8). The mean score for the Trust in Health Care Providers Subscale was 40.0 (SD 6.2), whereas the mean score for the Trust in Health Care Payers Subscale was 12.8 (SD 3.0), and the mean score for the Trust in Health Care Institutions Subscale was 10.3 (SD 2.1). The psychometric properties of each individual item and the total scale of the MTHCSS in the final sample are reported in Table 5. Mean total MTHCSS scores are reported by demographic characteristic in Table 6. Mean scores differed by gender (p = .04), race/ethnicity (p = .03), and having a usual source of care (p = .01). Trust scores were significantly lower among women, African Americans, and Hispanics and those without a usual source of care.
Table 5.
Items and Statistical Measures of the MTHCSS in Final Sample
| Item number | Wording of items | Means (SD) | Item selectivity | Cronbach’s a without item |
|---|---|---|---|---|
| Subscale 1: Trust in Health Care Providers (α = 0.92) | 40.0 (6.2) | |||
| 1 | My health care provider is usually considerate of my needs and puts them first | 4.3 (0.7) | 0.63 | 0.88 |
| 2 | I have so much trust in my health care provider that I always try to follow his/her advice | 4.2 (0.8) | 0.67 | 0.88 |
| 3 | I trust my health care provider so much that whatever he/she tells me, it must be true | 3.6 (1.0) | 0.65 | 0.88 |
| 4* | Sometimes, I do not trust my health care provider’s opinion and therefore I feel I need a second one | 3.2 (1.1) | 0.39 | 0.89 |
| 5 | I can trust my health care providers judgments concerning my medical care | 4.1 (0.7) | 0.65 | 0.88 |
| 6 | My health care provider will do whatever it takes to give me the medical care that I need | 4.1 (0.8) | 0.70 | 0.88 |
| 7 | Because my health care provider is an expert, he is able to treat medical problems like mine | 4.0 (0.7) | 0.64 | 0.88 |
| 8 | I can trust my health care provider’s decisions on which medical treatments are best for me | 4.1 (0.7) | 0.71 | 0.87 |
| 9 | My health care provider offers me the highest quality in medical care | 4.1 (0.7) | 0.70 | 0.88 |
| 10 | All things considered, I completely trust my health care provider | 4.1 (0.8) | 0.75 | 0.87 |
| Subscale 2: Trust in Health Care Payers (α = 0.74) | 12.8 (3.0) | |||
| 11 | Health care payers are good at what they do | 3.4 (0.9) | 0.40 | 0.89 |
| 12 | When needed, health care payers will pay for you to see any specialist | 3.1 (1.1) | 0.40 | 0.89 |
| 13 | When questioned about what treatments are covered, health care payers are honest with their answers | 3.3 (0.9) | 0.43 | 0.88 |
| 14 | Health care payers will pay for everything they are supposed to, including treatment that is expensive | 3.0 (1.1) | 0.35 | 0.89 |
| Subscale 3: Trust in Health Care Institutions (α = 0.64) | 10.3 (2.1) | . | ||
| 15* | Health care institutions only care about keeping medical costs down, and not what is needed for my health | 3.5 (1.0) | 0.25 | 0.89 |
| 16 | Healthcare institutions provide the highest quality in medical care | 3.5 (0.8) | 0.49 | 0.88 |
| 17 | When treating my medical problems, health care institutions put my medical needs above all other considerations, including costs | 3.4 (0.9) | 0.49 | 0.88 |
| Total scale (α = 0.89) | 63.0 (8.8) |
*Items 4 and 15 are reverse scored
Table 6.
Mean MTHCSS Scores by Sample Characteristics in the Final Sample
| Final sample (n = 301) | Mean (SD) | p values |
|---|---|---|
| Age | ||
| <35 years | 61.5 (8.9) | p = .19 |
| 35–49 years | 61.7 (9.5) | |
| 50–64 years | 63.6 (7.0) | |
| 65+ years | 64.5(10.1) | |
| Sex | ||
| Women | 62.1 (8.9) | p = .04* |
| Men | 64.3 (8.7) | |
| Race/Ethnicity | ||
| White | 64.3 (8.2) | p = .03* |
| African American | 61.5 (9.2) | |
| Hispanic/other | 61.4 (11.9) | |
| Marital status | ||
| Single | 62.1 (9.0) | p = .77 |
| Married | 62.9 (8.9) | |
| Separated/divorced | 63.2 (8.2) | |
| Widowed | 64.2 (9.8) | |
| Educational level | ||
| <High school graduate | 63.4 (10.7) | p = .94 |
| High school graduate | 62.8 (8.5) | |
| <College graduate | 63.2 (9.3) | |
| ≥College graduate | 62.5 (8.1) | |
| Insurance status | ||
| Private insurance | 62.6 (8.8) | p = .05 |
| Government insurance | 64.4 (9.2) | |
| Uninsured | 59.1 (6.5) | |
| Income | ||
| <$10,000 | 63.4 (8.2) | p = .41 |
| <$25,000 | 62.7 (9.1) | |
| <$35,000 | 63.3 (8.9) | |
| <$50,000 | 62.8 (6.8) | |
| <$75,000 | 61.1 (10.2) | |
| <$100,000 | 65.9 (7.5) | |
| $100,000+ | 62.6 (9.4) | |
| Usual source of care | ||
| Yes | 63.4 (8.4) | p = .01* |
| No | 59.7 (10.8) | |
| Saw health provider past 12 months | ||
| Yes | 63.1 (8.7) | p = .76 |
| No | 62.4 (9.5) | |
| Used Emergency room past 12 months | ||
| Yes | 63.2 (8.8) | p = .67 |
| No | 62.7 (8.9) | |
| Number of health provider visits in past 12 months | ||
| 0/1 | 60.8 (12.1) | p = .14 |
| 2 | 61.6 (7.2) | |
| 3 | 65.9 (6.8) | |
| 4 | 64.2 (9.3) | |
| 5+ | 63.6 (8.3) | |
| Number of emergency department visits in past 12 months | ||
| 0 | 61.9 (9.7) | p = .24 |
| 1 | 62.3 (8.7) | |
| 2 | 62.2 (7.7) | |
| 3 | 67.8 (7.2) | |
| 4+ | 65.9 (8.9) | |
*Statistically significant
Internal Consistency and Reliability Three factors (subscales) emerged from the confirmatory factor analysis (principal component analysis with Varimax Rotation) of the revised 17-item MTHCSS. The eigenvalues for the 3 components were 7.07, 2.26, and 1.10, respectively. These components of the scale explained 61% of the variance. The loading for the rotated items ranged between 0.36 and 0.86 (See Table 3). All individual items correlated at greater than or equal to .5 with at least 1 factor and did not correlate at greater than .3 with the other 2 factors29. Cronbach’s α for the total scale was 0.89 and 0.92, 0.74, and 0.64, respectively, for the 3 factors (subscales) in the final sample. The psychometric properties of each individual item and the total scale of the MTHCSS in the final sample are reported in Table 5.
MTHCSS Correlations with Trust Related Constructs Significant correlations were observed between the MTHCSS and measures of patient-centered care, locus of control, medication adherence, social support, and patient satisfaction (See Table 7). Correlations between the MTHCSS and the Modified Picker Scale30–32 (patient-centered care) ranged between r = −.22 and r = .62). A significant correlation between the MTHCSS was present on only 1 of the 3 Multidimensional Health Locus of Control Subscales33: Powerful Others (r = .42, p < .001). Significant correlations also existed between the MTHCSS and (1) the Morisky Scale of Medication NonAdherence34 (r = −.22, p < .001) and (2) the MOS Scale of Social Support35 (r = .253, p < .001). Finally, there were significant correlations between the MTHCSS total scale score and patient satisfaction (r = .67, p < .001) and dissatisfaction (r = .45, p < .001) with care.
Table 7.
Correlations Between the MTHCSS and Measures of Patient-centered Care, Health Locus of Control, Medication Adherence, Social Support, and Patient Satisfaction
| Measure | Correlation value |
|---|---|
| Patient-centered care—Modified Picker Scale | |
| Does your physician give you enough time to explain the reason for your visit? | 0.62* |
| When you ask questions, do you get answers that are understandable? | 0.51* |
| Does your physician take enough time to answer your questions? | 0.60* |
| Does your physician ask you about how your family or living situation might affect your health? | 0.45* |
| Do you get as much medical information as you want from your physician? | 0.67* |
| When you see your physician, do you have questions about your care that you want to discuss but do not? | −0.22* |
| Are you involved in decisions about your care as much as you want? | 0.47* |
| Health locus of control—Multidimensional Locus of Control Scale | |
| Internal (i.e., health/illness is determined by self) | 0.10 |
| Powerful others (i.e., health/illness is determined by others) | −0.04 |
| Chance (i.e., health/illness is determined by luck or chance) | 0.42* |
| Medication nonadherence—Morisky Scale | |
| Total scale | −0.22* |
| Social support—MOS Scale of Social Support | |
| Total scale | 0.25* |
| Satisfaction/dissatisfaction with care | |
| Looking over the past 12 months, I can say that I am very satisfied with the care received from my primary care provider | 0.67* |
| Looking over the past 12 months, I can say that I am very dissatisfied with the care received from my primary care provider | −0.45* |
*Statistically significant correlations at p < .05
DISCUSSION
We developed and tested a 17-item MTHCSS to examine the multiple objects of trust in the health care system. This scale includes items designed to measure trust in health care providers (10 items), trust in the health care payers (4 items), and trust in health care institutions (3 items). The psychometric properties of the MTHCSS indicate good reliability, validity, and acceptable psychometric properties. Similar psychometric properties were observed across 2 populations: college students and adults receiving care in primary care clinics. The MTHCSS has a high internal consistency (α = 0.89), suggesting that the 3 objects of trust can be captured in 1 scale.
The inclusion of 3 subscales is supported by the findings of Boulware et al.,25 which suggested that patterns of trust in physicians, health insurance plans, and hospitals can differ as a reflection of divergent experiences with the individual components of the health care system. Therefore, scales designed to measure trust should be multidimensional to account for the interplay between physicians, health insurance plans, and hospitals.
The MTHCSS was significantly correlated with established measures of patient-centered care, health locus of control, medication nonadherence, social support, and patient satisfaction. Scores on 5 of 6 items on the Modified Picker Scale30–32 showed significant correlations with patient trust. Patients who receive patient-centered care are more likely to have trust in the health care system. The final item, a negatively worded item regarding the patient’s ability to ask the physician questions during the visit, was also significantly correlated in a negative direction. Medication nonadherence (Morisky Scale)34 was negatively associated with trust indicating that trusting patients are more likely to take their medication as prescribed. Similarly, social support (MOS Scale of Social Support)35 and patient satisfaction were associated with trust. In summary, patients with more social support and satisfaction with their level of care were most likely to report greater trust.
Given that very few previously developed scales to measure trust capture the multiple objects of trust in the health care system concurrently,6–19 the multidimensional nature of this scale enables more accurate characterizations of trust in the health care system. This is especially important for studies of trust/distrust between racial/ethnic groups. The MTHCSS documented significant differences in trust across racial/ethnic groups with African Americans and Hispanics having lower levels of trust compared to whites. However, more studies are needed to validate this finding in diverse population groups and determine whether racial/ethnic differences in the different objects of trust influence health outcomes. The MTHCSS provides a valid and reliable tool for such studies.
This study has some limitations. First, the participants in the final study were recruited from 1 university clinic in the southeastern United States, and approximately 95% had health insurance. Therefore, it is uncertain how the scale will perform in other populations with different demographic characteristics. Additional studies are needed to examine the psychometric properties of the MTHCSS in diverse patient populations. Second, because of the small proportion of subjects in the final sample without a usual source of care (approximately 13%); the finding of lower levels of trust in patients without a usual source of care needs to be replicated.
In conclusion, the 17-item MTHCSS is a valid and reliable instrument to measure the multiple objects of trust in the health care system, and the psychometric properties of the scale were consistent across 2 diverse populations—college students and adults receiving care in primary care clinics.
Acknowledgment
We are grateful to the following summer undergraduate research students who assisted with data collection for the different phases of the project over the past 3 years: Douglas Gleaton, Melodie Harrison, Antoinette Bennett, Ashley Hambright, Sylvia Smith, and Brittany Smalls.
Conflict of interest None disclosed.
References
- 1.Hall MA, Dugan E, Zheng B, Mishra AK. Trust in physicians and medical institutions: what is it, can it be measured, and does it matter? Milbank Q. 2001;79(4):613–639. doi: 10.1111/1468-0009.00223. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Pearson SD, Raeke LH. Patients’ trust in physicians: many theories, few measures, and little data. J Gen Intern Med. 2000;15:509–513. doi: 10.1046/j.1525-1497.2000.11002.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Govier T.Distrust as a practical problem J Soc Philos 199223152–63. 10.1111/j.1467-9833.1992.tb00484.x11656183 [DOI] [Google Scholar]
- 4.Bigley GA, Pearce JL. Straining for shared meaning in organization science: problems of trust and distrust. Acad Manage Rev. 1998;23(3):405–422. doi: 10.2307/259286. [DOI] [Google Scholar]
- 5.Lewicki R, McAllister D, Bies R. Trust and distrust: new relationships and realities. Acad Manage Rev. 1998;23(3):438–458. doi: 10.2307/259288. [DOI] [Google Scholar]
- 6.Anderson LA, Dedrick RF. Development of the trust in physician scale: a measure to assess interpersonal trust in patient–physician relationships. Psychol Rep. 1990;67:1091–1100. doi: 10.2466/pr0.1990.67.3f.1091. [DOI] [PubMed] [Google Scholar]
- 7.Thom DH, Campbell B. Patient–physician trust: an exploratory study. J Fam Pract. 1997;44:169–176. [PubMed] [Google Scholar]
- 8.Kao AC, Green DC, Zaslavsky AM, Koplan JP, Cleary PD. The relationship between method of physician payment and patient trust. JAMA. 1998;280:1708–1714. doi: 10.1001/jama.280.19.1708. [DOI] [PubMed] [Google Scholar]
- 9.Mechanic D. Changing medical organization and the erosion of trust. Milbank Q. 1996;74:171–189. doi: 10.2307/3350245. [DOI] [PubMed] [Google Scholar]
- 10.Mechanic D, Meyer S. Concepts of trust among patients with serious illness. Soc Sci Med. 2000;51(5):657–668. doi: 10.1016/S0277-9536(00)00014-9. [DOI] [PubMed] [Google Scholar]
- 11.Thom DH, Ribisl KM, Stewart AL, Luke DA. Further validation and reliability testing of the Trust in Physician Scale. The Stanford Trust Study Physicians. Med Care. 1999;37(5):510–517. doi: 10.1097/00005650-199905000-00010. [DOI] [PubMed] [Google Scholar]
- 12.Hall MA, Zheng B, Dugan E, Camacho F, Kidd KE, Mishra A, Balkrishnan R. Measuring patients’ trust in their primary care providers. Med Care Res Rev. 2002;59(3):293–318. doi: 10.1177/1077558702059003004. [DOI] [PubMed] [Google Scholar]
- 13.Corbie-Smith G, Thomas SB, St George DM. Distrust, race, and research. Arch Intern Med. 2002;162(21):2458–2463. doi: 10.1001/archinte.162.21.2458. [DOI] [PubMed] [Google Scholar]
- 14.Hall MA, Camacho F, Dugan E, Balkrishnan R. Trust in the medical profession: conceptual and measurement issues. Health Serv Res. 2002;37(5):1419–1439. doi: 10.1111/1475-6773.01070. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Rose A, Peters N, Shea JA, Armstrong K. Development and testing of the health care system distrust scale. J Gen Intern Med. 2004;19(1):57–63. doi: 10.1111/j.1525-1497.2004.21146.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Thompson HS, Valdimarsdottir HB, Winkel G, Jandorf L, Redd W. The group-based medical mistrust scale: psychometric properties and association with breast cancer screening. Prev Med. 2004;38(2):209–218. doi: 10.1016/j.ypmed.2003.09.041. [DOI] [PubMed] [Google Scholar]
- 17.Goold SD, Fessler D, Moyer CA. A measure of trust in insurers. Health Serv Res. 2006;41(1):58–78. doi: 10.1111/j.1475-6773.2005.00456.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Zheng B, Hall MA, Dugan E, Kidd KE, Levine D. Development of a scale to measure patients’ trust in health insurers. Health Serv Res. 2002;37(1):187–202. [PubMed] [Google Scholar]
- 19.Dugan E, Trachtenberg F, Hall MA. Development of abbreviated measures to assess patient trust in a physician, a health insurer, and the medical profession. BMC Health Serv Res. 2005;5:64. doi: 10.1186/1472-6963-5-64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Doescher MP, Saver BG, Franks P, Fiscella K. Racial and ethnic disparities in perceptions of physician style and trust. Arch Fam Med. 2000;9(10):1156–1163. doi: 10.1001/archfami.9.10.1156. [DOI] [PubMed] [Google Scholar]
- 21.LaVeist TA, Nickerson KJ, Bowie JV. Attitudes about racism, medical mistrust, and satisfaction with care among African American and white cardiac patients. Med Care Res Rev. 2000;57(Suppl 1):146–161. doi: 10.1177/107755800773743637. [DOI] [PubMed] [Google Scholar]
- 22.Cort MA. Cultural mistrust and use of hospice care: challenges and remedies. J Palliat Med. 2004;7(1):63–71. doi: 10.1089/109662104322737269. [DOI] [PubMed] [Google Scholar]
- 23.Alston RJ. Racial identity and cultural mistrust among African-American recipients of rehabilitation services: an exploratory study. Int J Rehabil Res. 2003;26(4):289–95. doi: 10.1097/00004356-200312000-00006. [DOI] [PubMed] [Google Scholar]
- 24.Moseley KL, Freed GL, Bullard CM, Goold SD. Measuring African-American parents’ cultural mistrust while in a healthcare setting: a pilot study. J Natl Med Assoc. 2007;99(1):15–21. [PMC free article] [PubMed] [Google Scholar]
- 25.Boulware LE, Cooper LA, Ratner LE, LaVeist TA, Powe NR. Race and trust in the health care system. Public Health Rep. 2003;118(4):358–365. doi: 10.1016/S0033-3549(04)50262-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Laveist TA, Nuru-Jeter A. Is doctor–patient race concordance associated with greater satisfaction with care? J Health Soc Behav. 2002;43(3):296–306. doi: 10.2307/3090205. [DOI] [PubMed] [Google Scholar]
- 27.Cooper-Patrick L, Gallo JJ, Gonzales JJ, Vu HT, Powe NR, Nelson C, Ford DE. Race, gender, and partnership in the patient–physician relationship. JAMA. 1999;282(6):583–589. doi: 10.1001/jama.282.6.583. [DOI] [PubMed] [Google Scholar]
- 28.Statistical Package for the Social Sciences (SPSS 14.0). Chicago, IL: SPSS; 2006.
- 29.Nunnally JC, Bernstein I. Psychometric Theory. 3. New York: McGraw-Hill; 1994. [Google Scholar]
- 30.Cleary PD, Edgman-Levitan S, Roberts M, Moloney TW, McMullen W, Walker JD, Delbanco TL. Patients evaluate their hospital care: a national survey. Health Aff. 1991;10:254–267. doi: 10.1377/hlthaff.10.4.254. [DOI] [PubMed] [Google Scholar]
- 31.Cleary PD, Edgman-Levitan S, McMullen W, Delbanco TL. The relationship between reported problems and patient summary evaluations of hospital care. Qual Rev Bull. 1992;18:53–59. doi: 10.1016/s0097-5990(16)30507-3. [DOI] [PubMed] [Google Scholar]
- 32.Edgman-Levitan S, Cleary PD. What information do consumers want and need? Health Affairs. 1996;15(4):42–56. doi: 10.1377/hlthaff.15.4.42. [DOI] [PubMed] [Google Scholar]
- 33.Wallston KA, Wallson BS, DeVellis R. Development of the multidimensional health locus of control (MHLC) scales. Health Educ Mongr. 1978;6:160–170. doi: 10.1177/109019817800600107. [DOI] [PubMed] [Google Scholar]
- 34.Morisky DE, Green LW, Levine DM. Concurrent and predictive validity of a self-reported measure of medication adherence. Med Care. 1986;24(1):67–74. doi: 10.1097/00005650-198601000-00007. [DOI] [PubMed] [Google Scholar]
- 35.Sherbourne CD, Stewart AL. The MOS social support survey. Soc Sci Med. 1991;32(6):705–714. doi: 10.1016/0277-9536(91)90150-B. [DOI] [PubMed] [Google Scholar]