Search tips
Search criteria 


Logo of jgimedspringer.comThis journalToc AlertsSubmit OnlineOpen Choice
J Gen Intern Med. 2010 May; 25(5): 397–402.
Published online 2010 February 20. doi:  10.1007/s11606-010-1249-5
PMCID: PMC2854995

Patient and Physician Beliefs About Control over Health: Association of Symmetrical Beliefs with Medication Regimen Adherence



Past work suggests that the degree of similarity between patient and physician attitudes may be an important predictor of patient-centered outcomes.


To examine the extent to which patient and provider symmetry in health locus of control (HLOC) beliefs was associated with objectively derived medication refill adherence in patients with co-morbid diabetes mellitus (DM) and hypertension (HTN).


Eighteen primary care physicians at the VA Iowa City Medical Center and affiliated clinics; 246 patients of consented providers with co-morbid DM and HTN.


Established patient-physician dyads were classified into three groups according to the similarity of their HLOC scores (assessed in parallel). Data analysis utilized hierarchical linear modeling (HLM) to account for clustering of patients within physicians.


Objectively derived medication refill adherence was computed using data from the VA electronic pharmacy record; blood pressure and HgA1c values were considered as secondary outcomes.


Physician-patient dyads holding highly similar beliefs regarding the degree of personal control that individual patients have over health outcomes showed significantly higher overall and cardiovascular medication regimen adherence (p = 0.03) and lower diastolic blood pressure (p = 0.02) than in dyads in which the patient held a stronger belief in their own personal control than did their treating physician. Dyads in which patients held a weaker belief in their own personal control than did their treating physician did not differ significantly from symmetrical dyads. The same pattern was observed after adjustment for age, physician sex, and physician years of practice.


These data are the first to demonstrate the importance of attitudinal symmetry on an objective measure of medication adherence and suggest that a brief assessment of patient HLOC may be useful for tailoring the provider’s approach in the clinical encounter or for matching patients to physicians with similar attitudes towards care.

KEY WORDS: patient adherence, physician-patient interaction, health attitudes, chronic illness

It is well recognized that health care providers, as well as patients, differ broadly in regard to their health care-related attitudes, beliefs, and expectations16. For example, surveys have found that some patients prefer to take a highly active role during the clinical encounter, whereas others prefer to remain passive2,3,6. Similar differences have been observed with regard to physicians, with some holding more autocratic, physician-centered attitudes about clinical interactions, and others holding more egalitarian, patient-centered views1,46. However, the degree to which these attitudinal differences may be related to health care outcomes such as medical regimen adherence or to the effectiveness of clinical management is less clear.

Emerging evidence suggests that any attempt to consider the significance of patient or provider attitudes in isolation is insufficient (e.g.,1,47). Rather, this perspective suggests that a dyadic consideration of provider and patient attitudes or beliefs is essential to understanding their potential significance in influencing outcomes. Krupat was among the first to argue that the degree of symmetry between physician and patient attitudes may be an important consideration1,5,6. These investigators reported that patients who held similar attitudes to their physician (e.g., regarding preference for health-related information sharing) were more satisfied, more trusting, and more likely to recommend the physician to others.

More recently, we have examined whether the degree of patient-provider symmetry in the attitudes held toward health and health care delivery is associated with differences in patient adherence4. In one study, we expanded previous assessments of attitudinal differences to include measures of health locus of control (HLOC)8 beliefs. The HLOC construct has been one of the most actively studied attitudinal predictors of patients’ health-related behavior for decades9. Health locus of control is conceptualized as the degree to which a patient attributes the cause of health-related outcomes to internal factors under one’s own control (i.e., a patient’s own actions) or to external factors (e.g., chance, actions of the provider). Early work with this construct demonstrated that strong “internal” HLOC is associated with more positive adaptation to chronic disease when patient control over the illness or treatment is realistic, but may be maladaptive when there are impediments to exercising personal control10. In our previous study of physician-patient dyads4, we found that, among 146 patients seen by 16 different physicians, patients who were more similar in attitude to their providers, as indicated by HLOC scores, were more satisfied with care and self-reported better regimen adherence than patients whose control-related attitudes were less similar to those of their physicians.

Past studies involving the degree of symmetry between patient and provider attitudes have been limited in at least two important ways. First, earlier work has relied on self-ratings of patient adherence. It is well known that patients are often inaccurate in their ability to accurately self-report treatment adherence11. In the present study, we focused on objectively assessed medication refill adherence using the VA Medical Center’s electronic pharmacy record. Second, in our earlier study, we used a general, mixed convenience sample of primary care patients, and had no available information about diagnoses or treatment. Thus, the adherence assessment was not anchored to a specific regimen or condition, and the patient sample was highly heterogeneous. In the present study, we focused on patients with confirmed diagnoses of co-morbid diabetes mellitus (DM) and hypertension (HTN) treated in VA-affiliated outpatient primary care clinics.

The majority of patients with DM who are treated in VA facilities also have HTN, and increased emphasis has been placed on the need to control HTN in patients with DM12,13. Patients with DM are believed to get twice the benefit in cardiovascular risk reduction from HTN control compared to non-diabetics14, and patients with DM require more rigorous HTN control than patients without DM15,16. The population of patients with co-morbid DM and HTN is clearly a prevalent and clinically important group that has frequent contact with health care providers, whose conditions entail substantial self-management demands, and who are at high risk of increased morbidity and mortality.

In sum, the central aim of the study was to examine the extent to which patient and provider symmetry in HLOC beliefs was associated with the primary outcomes of medication refill adherence in patients with co-morbid DM and HTN. Blood pressure and glycemic control were secondary outcomes. Symmetry on the attitudinal HLOC measure was modeled before and after adjusting for patient and provider age, physician sex, patient income level, number of clinic visits between patient and provider during the 24-month index period, and physician years of practice1719.



Twenty-seven attending primary care physicians at the VA Iowa City Medical Center and affiliated clinics were contacted regarding the study. Most physicians consented to participate during the intermission of a VA-sponsored continuing medical education (CME) event. Physicians absent from the event were contacted and consented by mail and/or e-mail. Eighteen physicians (67%) consented and returned the study materials.

Patients were identified from appointment records of consented providers to create a dataset of patients grouped within providers. We contacted 517 consecutive patients with a diagnosis of co-morbid DM and HTN (based on ICD-9 codes from a prior clinic visit) seen in primary care clinics at the VA Iowa City Medical Center and affiliated clinics by a participating provider. Eligible patients had an active prescription for an anti-hypertensive medication and an oral hypoglycemic/anti-diabetic agent. Other inclusion criteria included having a home telephone and residing in an independent living environment. Patients were excluded if there was evidence of significant cognitive impairment as defined by a score of less than eight on the Short Portable Mental Status Questionnaire20. Three patients were excluded because of cognitive impairment. Three hundred (58%) of the eligible patients consented to participate. Only two patients were female, and these were excluded from analyses.

In an effort to choose patients who had an established relationship with their provider, patients had seen their provider for at least two appointments prior to the index appointment, within the past 24 months. Care was taken to ensure that each patient’s DM and HTN were managed by his/her primary care provider, as opposed to a specialty provider. Fifty-four patients were excluded from the sample because the index conditions were being managed or co-managed by an endocrinologist or cardiologist. Thus, the study sample consisted of 244 patients. The University of Iowa IRB and the VA Research and Development Committee approved all procedures. A summary of the samples is provided in Table 1.

Table 1
Characteristics of the Sample


Internal Health Locus of Control Scale, Form A8 The six-item Internal HLOC scale was used to assess beliefs about patients’ personal control over their health outcomes. Items were answered using a five-point Likert format with higher scores indicative of a stronger belief in patient control and a scoring range of 5–30 possible (see Table 2). For the traditional patient form of the measure, individuals responded to items assessing their beliefs and expectations concerning personal control over their own health. Physicians completed a previously adapted4 parallel version of the Internal HLOC scale, slightly revised to reflect a physician’s belief about patients’ level of personal control over health outcomes. For example, the original item “I am in control of my own health” became “Patients are in control of their own health.” An internal consistency alpha of 0.64 has been previously obtained for this adapted measure, and a full description of the measure and its reliability has been previously reported4.

Table 2
Internal Health Locus of Control Scale Items

Medication Refill Adherence Medication adherence was assessed by refill history for the 13 months prior to enrollment through a validated method that uses automated pharmacy data (i.e., medication possession ratio, MPR;21,22). This method assesses the timing of refills and infers the percentage of time patients have an undersupply of drug during the index period. We used refill data for all regularly scheduled cardiac and oral diabetes medications prescribed by the participating physician in the preceding 13 months. The nonadherence MPR was calculated as: equation M1. Separate MPR nonadherence scores were computed for cardiovascular medications (CV; e.g., lisinopril, hydrochlorithiazide, pravastatin) and for oral diabetic medications (e.g., metformin, glyburide, rosiglitazone) across all agents taken by the patient and divided by the number of drugs in order to obtain a single, continuous mean MPR score for that category. In addition, an overall MPR was computed reflecting average nonadherence across all prescribed drugs in both categories.

Glycemic Control The mean of the available hemoglobin A1c laboratory values from all VA outpatient visits, assessed over 18 months prior to study enrollment, was used as an index of glycemic control. HgA1c provides a measure of the average blood glucose over the prior 2–3 months. An average of 2.0 HgA1c values (range 1–4) was available for each patient.

Blood Pressure (BP) The mean of the available systolic/diastolic BP readings obtained in all outpatient clinic visits was assessed over the 18 months prior to study enrollment. An average of 5.1 BP values (range 1–20) was available for each patient. Values were obtained from the VA electronic medical record.


Participating physicians completed survey materials at the intermission of a CME event or returned them by mail at a later date. Eligible patients were mailed a packet of baseline questionnaires, including the informed consent documents and HLOC measure 2 weeks prior to their scheduled primary care appointment with the participating physician. One week prior to the appointment, patients received a telephone call from a research assistant who reviewed the consent documents, outlined study procedures, and answered questions. If interested, patients were administered the mental status screen and, if eligible, asked to provide written informed consent and complete the questionnaires. At the patient’s scheduled appointment, a research assistant greeted the patient, received the survey, and gathered additional information concerning the clinic visit following the conclusion of their visit. The focus of the current report is on the pre-visit assessment of patient and physician attitudes; data obtained in the post-visit assessment have been reported elsewhere23.


To examine the effect of patient-physician symmetry with respect to HLOC beliefs on outcomes, patient-physician dyads were classified into three symmetry groups according to similarity of their HLOC scores. Dyads having the lowest patient-physician difference were assigned to the “symmetrical” group, with the defining cutoff difference chosen such that approximately one-third of the dyads would be in this group. The remaining dyads were assigned to the “patient high/physician low” group if the patient HLOC score exceeded that of the physician (reflecting a stronger belief in patient control over health) or to the “patient low/physician high” group if the physician HLOC score exceeded that of the patient. Data analysis utilized hierarchical linear modeling (HLM;24) to account for clustering of patients within physicians. Both patient and physician were treated as random factors. HLM analyses were used to examine the extent to which patient and physician symmetry in HLOC beliefs were associated with our primary outcome and medication refill adherence (cardiovascular medications; diabetes medications; all medications) as well as our secondary clinical outcomes (BP, HgA1c). Symmetry on the attitudinal measure was an independent variable in all models. Covariates in the adjusted models included patient income, age, and number of visits, and physician age, sex, and years of practice. We also performed follow-up pairwise comparisons, if the corresponding global test for a given dependent variable was significant, using the Tukey multiple comparison adjustment for p-values. Tests were performed using the SAS PROC MIXED. All tests were two-tailed with alpha = 0.05.


Table 1 shows the sample characteristics. Fifty-four patients were excluded because index conditions were being managed or co-managed by another provider. Thirteen patients failed to complete any part of the survey, while seven were missing some or all of the HLOC and were also excluded, resulting in a final sample size of 224 dyads. Inspection of a scatter plot of the patient and physician HLOC scores showed that the patient scores had a similar distribution across physicians. Table 3 shows the number of dyads in each of the three symmetry groups and the corresponding cutpoints that define these groups.

Table 3
Symmetry Category Dyad Frequencies and Corresponding Dyad-Difference Ranges

Table 4 shows the covariate-adjusted means and both unadjusted and adjusted overall model test results. Global adjusted tests showed a HLOC symmetry effect for overall MPR (F2,202 = 4.13, p = 0.02), CV MPR (F2,202 = 3.86, p = 0.02), and diastolic blood pressure (F2,164 = 3.24, p = 0.04), but not for diabetic MPR (F2,142 = 1.93, p = 0.15). Follow-up pairwise comparisons showed the “symmetrical” group to have significantly higher overall medication adherence for overall MPR and CV MPR and lower diastolic blood pressure compared to the “patient high/physician low” HLOC group, but not compared to the patient low/physician high group. Patient MPRs for the symmetrical dyads evidenced an overall average absence of medication on 12% of prescribed days over the 13-month assessment period compared to 18% of prescribed days among the dyads in which patients held higher control beliefs than their physicians, the latter reflecting a 50% higher rate of refill nonadherence. Although the global and follow-up tests for oral diabetic MPR did not attain significance, a similar trend existed, with the "symmetrical" group having higher medication adherence compared to the "patient high/physician low" HLOC group (Tukey pairwise p = 0.136). For a difference of this magnitude (0.14 vs. 0.09), posthoc power computation showed the power to be less than 0.50. Unadjusted results mirrored those described above. We note that patient age was the only significant covariate in the adjusted models reflecting better medication adherence, better glycemic control, and lower diastolic blood pressure with advancing age.

Table 4
Adjusted Means and Adjusted and Unadjusted Test Results


The present study builds on previous research and theory1,4, suggesting that similarity in health-related attitudes held by the physician and patient is a significant correlate of patient-centered outcomes such as medication regimen adherence. The present data are the first to demonstrate the importance of attitudinal symmetry on a measure of adherence that was not reliant on patient self-reports and that was anchored to a specific regimen and medical condition. Specifically, in physician-patient dyads holding highly similar beliefs regarding the degree of personal control that individual patients have over health outcomes, both overall and cardiovascular medication regimen adherence was significantly higher than in dyads in which the patient held a stronger belief in their own personal control than their treating physician’s beliefs regarding patients in general. A similar pattern was observed for diastolic, but not for systolic, blood pressure. This pattern was observed in models both adjusted and unadjusted for several patient and physician characteristics. These statistically significant differences in medication adherence appear to be clinically significant as well given the approximately 50% higher rate of nonadherence in the asymmetrical group, which corresponds to an absolute group difference of approximately 0.50 SD25,26. The significant group differences in diastolic blood pressure appear less clinically meaningful. Significant symmetry effects were not observed for oral diabetic medication adherence or for HgA1c values.

Interestingly, it was the subgroup of physician-patient dyads marked by high patient and low physician control beliefs that demonstrated poorer medication regimen adherence on two of three such metrics. That is, patients with stronger health-related control expectations appear to show significantly poorer adherence when managed by providers who hold lower expectations about the degree of control that patients can conceivably assert over health outcomes. This pattern is consistent with previous psychological theory that individuals who value or expect control in an important domain display increased negative affect and are more likely to engage in maladaptive behavioral responses when control is threatened or constrained in some way27. We have previously suggested that patients’ attempts to reassert control when they perceive it is threatened—either by the objective constraints of a clinical situation or by their perceptions of the clinical context—may be an important explanatory mechanism for some cases of nonadherence23,28. Although speculative, one interpretation is that patients with high control beliefs displayed poorer adherence when interacting with providers holding more skeptical beliefs about patient control over health due, in part, to a perception of restricted opportunities to exercise control.

These data have potentially important relevance to clinical practice. First, the current results suggest that the feasibility of matching patients and providers based upon similarity on the HLOC attitudinal measure should be examined. Whether such matching is done explicitly or whether patients are informed which available providers have similar attitudes and are allowed to make their own selection (e.g.,29), these data suggest that patients who are “matched” with providers holding similar attitudes will exhibit better medication regimen adherence. An alternative approach could involve instructing physicians in tailoring their practice style to better fit an individual patient’s attitudes or expectations. Evidence exists to suggest that providers do adapt their practice style as a function of certain patient background characteristics17,3032. Our belief, however, is that the key element in any approach to tailoring the physician-patient interaction is the availability of readily obtained information about patient attitudes or preferences. The current results suggest that the brief, HLOC instrument provides a snapshot of a patient’s health-related attitudes that could guide provider efforts to tailor their approach to a particular patient’s orientation. Such real-time access to information regarding patient preferences and expectations could, for example, be used to more strongly engage selected patients in treatment decisions and self-management processes while offering less patient involvement and greater provider control with other patients.

These results should be tempered by certain study limitations. Physicians and patients were sampled from a single VA site, and it is possible that the resulting sample of older male, predominantly Caucasian, veterans may not be representative of patients in other settings. The pattern of results involving attitudes and medication adherence may differ in certain ethnic minority patient groups or in female patients. Moreover, while the response rate among eligible physicians was fairly high (74%), it was lower (58%) among eligible patients, and it is possible that the final samples were not representative of their larger populations, making it necessary to use caution in generalizing these results. Our assessment of patients’ medication adherence was limited to a single method based on automated pharmacy records within the VA. Although the use of refill records to assess adherence has become increasingly common (e.g.,33), this method does not ensure that a patient is taking the drug as directed. Moreover, non-VA medications may not be as reliably accounted for in the available electronic VA record as are VA prescribed and filled prescriptions. It is also the case that the calculation of MPR values may not account for some “mid-course” changes to a regimen. For example, if a drug was abruptly discontinued by the provider, the MPR value for the drug may be artificially reduced by the number of days remaining on the prescription. For these reasons the MPR metric should be interpreted with appropriate caution. Nevertheless, evidence firmly supports the validity of the MPR method with strong associations reported between pharmacy records and other measures of adherence and with drug serum levels and clinical effects3436. A recent review of adherence assessment in hypertension concluded that the MPR is the “best available measure” for assessing adherence using retrospective data34. Perhaps most importantly, although the MPR method has limitations, there is no reason to believe that these limitations reflect a systematic bias that differs across our attitudinal subgroups.

Finally, while effects were found on measures representing adherence to both cardiovascular and overall medication regimens, an effect was not found on adherence to oral diabetic medications. Note, however, that the pattern of differences observed for adherence to diabetic medications paralleled that seen for the other outcomes, and a posthoc power computation showed low power (<0.50), suggesting that the lack of an effect may have been a function of limited statistical power. These limitations notwithstanding, the present data suggest that symmetry between patient and provider beliefs about health-related control is associated with patient adherence. Further work is needed to directly examine whether an assessment of patient beliefs about control over health may be a useful tool for enhancing adherence when used either as a guide for tailoring the approach that the provider takes in the clinical encounter or when used to match patients to physicians with similar attitudes towards care.


The research reported here was supported in part by the Department of Veterans Affairs, Veterans Health Administration, Health Services Research and Development Service grant no. IIR04-201 awarded to Dr. Christensen, who is a Senior Scientist in the Center for Research in the Implementation of Innovative Strategies in Practice (CRIISP) in the VA-Iowa City Healthcare System. The views expressed are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs. This project was also supported in part by an Agency for Healthcare Research and Quality (AHRQ) Centers for Education and Research on Therapeutics cooperative agreement no. 5 U18 HSO16094.

There are no potential conflicts of interest for any author.


The views expressed are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs.


1. Krupat E, Yeager CM, Putnam S. Patient role orientations, doctor-patient fit, and visit satisfaction. Psychol Health. 2000;15:707–719. doi: 10.1080/08870440008405481. [Cross Ref]
2. Levinson W, Kao A, Kuby A, Thisted RA. Not all patients want to participate in decision making: A national study of public preferences. J Gen Intern Med. 2005;20:531–535. doi: 10.1111/j.1525-1497.2005.04101.x. [PMC free article] [PubMed] [Cross Ref]
3. Swenson SL, Buell S, Zettler P, et al. Patient centered communication: Do patients really prefer it? J Gen Intern Med. 2004;19:1069–1079. doi: 10.1111/j.1525-1497.2004.30384.x. [PMC free article] [PubMed] [Cross Ref]
4. Cvengros JA, Christensen AJ, Hillis SL, Rosenthal GE. Patient and physician attitudes in the healthcare context: Attitudinal symmetry predicts patient satisfaction and adherence. Ann Behav Med. 2007;33:262–268. doi: 10.1007/BF02879908. [PubMed] [Cross Ref]
5. Krupat E, Bell RA, Kravitz RL, et al. When physicians and patients think alike: patient-centered beliefs and their impact on satisfaction and trust. J Fam Pract. 2001;50:1057–1062. [PubMed]
6. Krupat E, Rosenkranz SL, Yeager CM, et al. The practice orientations of physicians and patients: Effect of doctor-patient congruence on satisfaction. Patient Educ Couns. 2000;39:49–59. doi: 10.1016/S0738-3991(99)00090-7. [PubMed] [Cross Ref]
7. Jhang KH, Martin LR, Golin CE, DiMAtteo MR. Preferences for medical collaboration: Patient-physician congruence and patient outcomes. Patient Educ Couns. 2005;57:308–314. doi: 10.1016/j.pec.2004.08.006. [PubMed] [Cross Ref]
8. Wallston KA, Wallston BS, DeVellis R. Development of the Multidimensional Health Locus of Control (MHLC) Scales. Health Educ Monogr. 1978;6:160–170. [PubMed]
9. Wallston KA. Overview of the special issue on research with the Multidimensional Health Locus of Control (MHLC) Scales. J Health Psychol. 2005;10:619–621. doi: 10.1177/1359105305055301. [Cross Ref]
10. Christensen AJ, Turner CW, Smith TW, et al. Health locus of control and depression in end-stage renal disease. J Consult Clin Psychol. 1991;59:419–424. doi: 10.1037/0022-006X.59.3.419. [PubMed] [Cross Ref]
11. Wang PS, Bener JS, Glynn RJ, et al. How well do patients report noncompliance with antihypertensive medications? A comparison of self report versus filled prescriptions. Pharmacoepidemiol Drug Saf. 2004;13:11–19. doi: 10.1002/pds.819. [PubMed] [Cross Ref]
12. Mooradian AD. Cardiovascular disease in type 2 diabetes mellitus: Current managment guidelines. Arch Intern Med. 2003;163:33–40. doi: 10.1001/archinte.163.1.33. [PubMed] [Cross Ref]
13. Vijan S, Hayward RA. Treatment of hypertension in type 2 diabetes mellitus: Blood pressure goals, choice of agents, and setting priorities in diabetes care. Ann Intern Med. 2003;138:593–602. [PubMed]
14. Curb JD, Pressel SL, Cutler JA, et al. Effect of diuretic-based antihypertensive treatment on cardiovascular disease risk in older patients with isolated systolic hypertension. Systolic Hypertension in the Elderly Program Cooperative Research Group. JAMA. 1996;276:1886–1892. doi: 10.1001/jama.276.23.1886. [PubMed] [Cross Ref]
15. UK Prospective Diabetes Study Group Tight blood pressure control and risk of macrovascular and microvascular complications in type 2 diabetes: UKPDS 38. Br Med J. 1998;317:703–713. [PMC free article] [PubMed]
16. Kjeldsen SE, Hedner T, Jamerson K, et al. Hypertension optimal treatment (HOT) study: Home blood pressure in treated hypertensive subjects. Hypertension. 1998;31:1014–1020. [PubMed]
17. Makoul G. Perpetuating passivity: Reliance and reciprocal determinism in physician-patient interaction. J Health Comm. 1998;3:233–259. doi: 10.1080/108107398127355. [PubMed] [Cross Ref]
18. Callahan EJ, Bertakis KD, Azari R, et al. The influence of patient age on primary care resident physician-patient interaction. J Am Geriatr Soc. 2000;48:30–35. [PubMed]
19. Bertakis KD, Azari R, Callahan EJ, Robbins JA, Helms LJ. Comparison of primary care resident physicians’ practice styles during initial and return patient visits. J Gen Intern Med. 1999;14:495–498. doi: 10.1046/j.1525-1497.1999.05068.x. [PMC free article] [PubMed] [Cross Ref]
20. Pfeiffer E. A short portable mental status questionnaire for the assessment of organic brain deficit in elderly patients. J Am Geriatr Soc. 1975;23:433–441. [PubMed]
21. Steiner JF, Koepsell TD. Fihn, SD, Inui TS. A general method of compliance assessment using centralized pharmacy records. Med Care. 1988;26:814–823. doi: 10.1097/00005650-198808000-00007. [PubMed] [Cross Ref]
22. Billups SJ, Malone DC, Carter BL. The relationship between drug therapy noncompliance and patient characteristics, health-related quality of life, and health care costs. Pharmacotherapy. 2000;20:941–949. doi: 10.1592/phco.20.11.941.35266. [PubMed] [Cross Ref]
23. Cvengros JA, Christensen AJ, Hillis SH, Kaboli, P. Patient preference for and reports of provider behavior: Impact of symmetry on adherence, satisfaction, and glycemic control. Health Psychol 2009 (In press). [PubMed]
24. Raudenbush SW, Bryk AS. Hierarchical linear models: Applications and data analysis methods. 2. Thousand Oaks: Sage; 2002.
25. Norman GR, Sloan JA, Wyrwich KW. Interpretation of changes in health-related quality of life: The remarkable universality of half a standard deviation. Medical Care. 2003;5:582–592. doi: 10.1097/00005650-200305000-00004. [PubMed] [Cross Ref]
26. Sloan JA, Dueck A. Issues for statisticians in conducting analyses and translating results for quality of life endpoints in clinical trials. J Biopharm Stat. 2004;14:73–96. doi: 10.1081/BIP-120028507. [PubMed] [Cross Ref]
27. Brehm JW. The theory of psychological reactance. New York: Wiley; 1996.
28. Christensen AJ. Patient adherence to medical treatment regimens. Yale University Press; 2004.
29. Brink-Muinen A, Dulmen S, Messerli-Rohrbach V, Bensing J. Do gender-dyads have different communication patterns? A comparative study in Western-European general practices. Patient Educ Couns. 2002;48:253–264. doi: 10.1016/S0738-3991(02)00178-7. [PubMed] [Cross Ref]
30. Callahan EJ, Bertakis KD, Azari R, et al. The influence of patient age on primary care resident physician-patient interactions. J Am Geriatr Soc. 2000;48:30–35. [PubMed]
31. Krupat E, Hsu J, Irish J, et al. Matching patients and practitioners based on beliefs about care: Results of a randomized controlled trial. Am J Manag Care. 2004;10:814–822. [PubMed]
32. Lewin SA,Skea Z, Entwistle VA, et al. Interventions for providers to promote a patient-centered approach in clinical consultations. Cochrance Database Syst Rev 2001;4:Art.No.:CD003267. doi: 10.1002/14651858.CD003267. [PubMed]
33. Laine C, Newschaffer CJ, Zhang D, et al. Adherence to antiretroviral therapy by pregnant women infected with human immunodeficiency virus: a pharmacy claims-based analysis. Obstet Gynecol. 2000;95:167–173. doi: 10.1016/S0029-7844(99)00523-2. [PubMed] [Cross Ref]
34. Saunders K, Simon G, Bush T, Grothaus L. Assessing the feasibility of using computerized pharmacy refill data to monitor antidepressant treatment on a population basis: A comparison of automated and self-report data. J Clin Epidemiol. 1998;51:883–890. doi: 10.1016/S0895-4356(98)00053-5. [PubMed] [Cross Ref]
35. Hansen RA, Kim MM, Song L, et al. Comparison of methods to assess medication adherence and classify nonadherence. Ann Pharmacother. 2009;43:413–422. doi: 10.1345/aph.1L496. [PubMed] [Cross Ref]
36. Halpern MT, Khan ZM, Schimer JK, et al. Recommendations for evaluating compliance and persistence with hypertension therapy using retrospective data. Hypertension. 2006;47:1039–1048. doi: 10.1161/01.HYP.0000222373.59104.3d. [PubMed] [Cross Ref]

Articles from Journal of General Internal Medicine are provided here courtesy of Society of General Internal Medicine