|Home | About | Journals | Submit | Contact Us | Français|
Objective To determine if patients vary in perceptions of safety if interpersonal continuity were to be disrupted. If so, which characteristics are associated with feeling unsafe?
Background The extent to which patients’ preference for continuity with a personal physician is due to perceptions of safety is unclear.
Design Observational study (Wisconsin Longitudinal Study Graduate and Sibling Survey).
Setting and participants A total of 6827 respondents (most aged 63–66years) who completed the 2003–06 survey round.
Main variables studied Age, gender, marital status, education, health insurance type, illnesses, medications, length of relationship with provider and place, personality type, decision‐making preference and trust in physician deliberation.
Main outcome measures Safety perception when visiting another doctor or clinic if own doctor were not available.
Results Twelve percent of respondents felt unsafe. After adjustment, as compared to those who felt safe, those who felt unsafe were more likely to be women (Odds ratio=1.65, 95% confidence interval=1.35–2.01), have more chronic conditions (1.27, 1.08–1.50) and have a longer relationship with a usual provider: 5–9years (1.53, 1.11–2.10) 10–14years (1.41, 1.02–1.95) and 15 or more years (1.62, 1.20–2.17) compared to 0–4years. Those who preferred active participation in decision making and had trust in their physician were less likely to feel safe (1.63, 1.10–2.41).
Conclusions Certain older adults perceive being unsafe if not seeing their usual physician. Further research should investigate reasons for perceptions of safety if continuity were disrupted and any implications for care.
Interpersonal continuity of care between a patient and a personal physician is positively associated with several patient‐centered outcomes including satisfaction, 1 , 2 , 3 , 4 trust 5 , 6 and adherence. 7 However, patients vary in their preference for interpersonal continuity. 8 In particular, seniors, 9 , 10 females, 11 the less educated, medicare/medicaid recipients, 10 those with more chronic conditions 11 and those with longer relationships to their physician 10 , 12 prefer interpersonal continuity. Specific reasons given for this preference include a perception of comfort 13 or familiarity with the physician, 14 , 15 better communication 16 , 17 and the physician’s prior knowledge about them. 14 , 15
It has been suggested that discontinuity in care can lead to decreased patient safety by inhibiting a clinician’s ability to anticipate and correct gaps in care. 18 However, it is unclear the extent to which patients’ interpersonal continuity preference is due to perceptions that they are safer seeing the same physician. If a preference for interpersonal continuity is related to perceptions of safety, discontinuity with a physician may have severe consequences such as decreased adherence in addition to the safety concerns cited above. Learning about patients who feel unsafe when seeing physicians other than their own is increasingly important given the health system shift away from interpersonal continuity towards continuity provided by a team of physicians at a usual site 19 or insurance changes that lead to disruptions in care. 20
Patients’ perceptions of safety also are important because these perceptions differ from those of health‐care professionals and may have consequences for care. Although information regarding patients’ perceptions of safety is limited, 21 , 22 , 23 , 24 studies suggest that patients conceive of safety broadly by considering areas such as access, care coordination and communication in addition to technical quality. 23 , 25 , 26 , 27 A 1997 Harris poll of 1513 randomly selected households in the United States found that 42% of adults reported being personally involved in a situation where a medical mistake was made. One‐third of those adults further believed that the medical mistake had a permanent negative effect on health. 26 Patients describe physical, emotional and financial harm occurring from medical mistakes. 28 Some patients who perceive experiencing a medical error change future behaviour by taking extra measures to ensure their own safety. 24 , 26 , 29 Others report then mistrusting the health‐care system and avoiding further interaction, or seeking out another provider. 24 , 28
This paper seeks to answer the following questions: (1) Is there variation in the extent to which patients feel unsafe if interpersonal continuity were to be disrupted? (2) If there is variation, which patient characteristics are associated with feeling unsafe? In addition to previous factors associated with general preference for continuity (e.g. age, gender, education, insurance, chronic conditions, length of patient‐physician relationship), we examine the patient’s length of relationship with their usual place of care as well as their decision‐making preferences and personality characteristics. We also examine whether the effect of these characteristics was mediated by trust in the usual physician.
We use data from the Wisconsin Longitudinal Study (WLS) survey, a long‐term cohort study of a one‐third random sample (n=10317) of men and women who graduated from Wisconsin high schools in the spring of 1957 and 7638 of their randomly‐selected brothers and sisters. Individuals were most recently surveyed in 1992–94 and 2003–06. In the 2003–06 round of data collection, all surviving WLS participants were contacted via telephone whether or not they participated in the 1992 round of the survey. Respondents were traced and interviews were conducted and audio‐recorded using computer‐assisted techniques by the University of Wisconsin Survey Center. Interviews lasted approximately 1h. Among graduate survivors, the response rate for this survey was 80% and for siblings the response rate was 78%. Telephone interviews were followed by 54‐page mail‐out, mail‐back surveys that take approximately 90min to complete. Three mailings and one final telephone contact were made to encourage respondents to mail back their questionnaires. Among those who completed the telephone interview, a response rate of 88% was achieved for graduates and 81% for siblings.
We included graduate and sibling respondents who answered the question about safety. We excluded 8% of respondents who did not perceive they had a usual health‐care provider and therefore could not have interpersonal continuity. An additional 10% of respondents who were unsure of the length of relationship that they had with their provider were also excluded. The final sample size was 6827. This study was approved by the Institutional Review Board at the participating university.
All variables are obtained through patient self‐report. The degree to which patients feel safe if interpersonal continuity were to be disrupted, the dependent variable, was indicated by response to the following item: ‘If my doctor were not available, I would feel safe visiting another doctor or clinic’ (1=agree strongly to 5=disagree strongly).
Socio‐demographic information included age, gender, marital status, education and health insurance type. Other variables relating to health and physician–patient relationship included chronic illness measured by the Duke Older Americans Resources and Services (OARS) schedule of common illnesses and conditions 30 plus several additional chronic illnesses and conditions, the number of regularly taken prescription medications, the length of relationship with a usual provider and the length of relationship with a usual place of care. For conditions other than hypertension, diabetes, cancer (other than skin), coronary artery disease, stroke, joint pain and chronic arthritis, information was collected on the extent to which the condition affected the respondent’s life (1=Not at all to 5=A great deal). A variable was constructed that summed the number of conditions that affected the respondent’s life ‘quite a bit’ or ‘a great deal’.
Decision‐making preference was constructed from four variables that assessed respondent preferences during health care visits for information exchange (physician knowledge of patient medical history and physician disclosure of treatment choices), deliberation (discussion of treatment choices) and selection of treatment choice. Cluster analysis was performed on these items to create a typology of patient preference types. 31 Four preference types characterized over 96% of respondents, all of which preferred high levels of information exchange with physicians. Deliberative autonomists were the most active type; they preferred discussion of treatment choices and personal control over important decisions. Deliberative delegators preferred discussion of treatment choices with doctor control over important decisions. Non‐deliberative autonomists preferred little or no discussion of treatment choices with personal control over important decisions. Non‐deliberative delegators were the least active type; they preferred little or no discussion of treatment choices and doctor control over important decisions.
We measured personality based on the widely‐used five factor model. 32 Twenty‐nine items—a subset of the BFI‐54 33 —represented the five‐factors: extraversion (six items), agreeableness (six items), conscientiousness (six items), neuroticism (five items) and openness‐to‐experience (six items) with response categories measured on a 6‐point scale from 1=agree strongly to 6=disagree strongly. Items within each factor were combined into a scale, and all scales were then standardized to have a mean of zero and standard deviation of one.
Patient’s trust in physician deliberation was assessed using responses to the question, ‘If I had many treatment options, I worry about whether my doctor cares enough to discuss each one with me as long as I want’ (1=agree strongly, 5=disagree strongly; disagree strongly/disagree=high trust, neutral=neutral and agree strongly/agree=low trust).
Logistic regression was used to estimate odds ratios and 95% confidence intervals. We compared individuals who would feel unsafe with disruption to interpersonal continuity (response of agree strongly/agree) to those who would feel safe if interpersonal continuity were disrupted (response of disagree strongly/disagree) for each explanatory variable (socio‐demographics, chronic health conditions, count of life‐interfering conditions, a count of medications regularly taken, length of relationship with usual provider/place, decision‐making preference type and personality factors), controlling for all other variables in the model.
Because we expected that the effect of these characteristics on feeling unsafe might differ by trust, we tested for interaction between trust in physician deliberation and the following: gender, marital status, personality and decision making preferences. We accounted for clustering of siblings within families by calculating confidence intervals and significance tests using the Stata ‘robust’ command, clustering family explicitly. 34 , 35 , 36 Results were considered statistically significant at P‐value<0.05. Likelihood ratio tests were conducted to determine the statistical significance for groups of indicator variables. Data were analysed using SAS 9.1 and Stata 9.0. 37 , 38
Respondents were predominantly in their early to mid sixties (Table 1). Most were married with at least a high school education and held private insurance. The majority had longstanding relationships with their usual place and provider of over 5years. Hypertension, joint pain and arthritis were common conditions. Diabetes, cancer, coronary artery disease and stroke were less common conditions present in the sample. Over half of the respondents were taking at least one prescription medication regularly. Less than 20%, however, had other chronic conditions that interfered quite a bit or greatly with their daily lives. The majority of respondents were deliberators and therefore preferred discussion of treatment options with their physician.
Variation existed in the extent to which patients felt unsafe were they to experience discontinuity. Figure 1 shows the percentage distribution of this variation; 67% of the sample agreed or strongly agreed that they would feel safe visiting someone other than their own doctor (n=4555), 21% were neutral (n=1404) and 12% disagreed or strongly disagreed with this statement (n=868).
Table 2 indicates characteristics of individuals who felt unsafe (n=868) as opposed to safe (n=4555) about visiting another doctor or clinic. Patients who feel unsafe were more likely to be female (Odds Ratio=1.65; 95% Confidence Interval: 1.35–2.01). Age, education, marital status and insurance type were not significantly associated with feeling unsafe.
The length of relationship with a provider but not a place was significantly associated with feeling unsafe versus safe with a break in interpersonal continuity. As compared to those who had been with their provider 0–4years, those who had been with their provider for 5–9years had 53% increased odds of feeling unsafe (95% CI: 1.11–2.10), those who had been with their provider10–14years had 41% increased odds of feeling unsafe (95% CI: 1.02–1.95) and those who had been with their provider 15 or more years had 62% increased odds of feeling unsafe (95% CI: 1.20–2.17).
While having an increasing number of chronic conditions that interfered with life was significantly associated with feeling unsafe with a break in interpersonal continuity (OR=1.27; 95% CI: 1.08–1.50), only one specific chronic condition, coronary artery disease, was significantly associated with feeling unsafe (OR=1.49, 95% CI: 1.18–1.88). As a sensitivity analysis, we repeated analyses with a variable that indicated the presence of any chronic condition. This variable was not significantly associated with feeling unsafe. Taking more medications, decision‐making preference type and personality type also were not significantly associated with safety.
Individuals who felt safe or unsafe about seeing someone other than their own physician varied in trusting their physician’s deliberation. Of those who felt safe about seeing someone else, 49% perceived high trust, 22% were neutral and 29% had low trust (data not shown). Conversely, of those who felt unsafe about seeing someone else, 55% perceived high trust, 15% were neutral and 30% perceived low trust. There was a borderline statistically significant interaction between decision making preference and trust (P<0.06), suggesting that the effect of decision‐making preference on feeling unsafe about a break in continuity might differ for patients who had weak versus high trust in their physician. As indicated in Table 3, among those with high trust in their physician, deliberative autonomists (the most active preference type) were more likely to feel unsafe if there were to experience discontinuity when compared to non‐deliberative delegators (the least active preference type) (OR 1.63; 95% CI: 1.10–2.41). A similar result was shown for deliberative delegators (OR 1.50; 95% CI: 0.97–2.32) although this result was only borderline statistically significant.
To our knowledge, this is the first paper that examines patients’ sense of safety if they were to experience a break in interpersonal continuity. Our study supports the idea that interpersonal continuity is preferred by those in vulnerable subgroups 8 , 10 as we found that women, patients with more chronic conditions and patients with coronary artery disease would feel more unsafe if seeing someone other than their own doctor. Overall, decision‐making preference and personality were not associated with feeling unsafe, although those with a preference for active participation in decision making (i.e. deliberative autonomists) were more likely to feel unsafe if they had high trust in their physician’s deliberation.
While other studies have indicated that a majority of individuals prefer interpersonal continuity, 11 , 39 we found that only 12% of the predominantly older adults in our study reported that they would feel unsafe if continuity were to be disrupted. This finding raises the question of whether, while interpersonal continuity is an ideal preferred by many individuals, actual perceptions of consequences due to discontinuity may be evident in only a smaller group of individuals. This magnitude in difference between individuals who prefer continuity and who actually perceive the consequence of feeling unsafe merits further study, particularly as interpersonal continuity is studied for tangible links to health benefit or cost reduction.
Although our study suggests that some individuals perceive disruption to continuity as unsafe, there is evidence that interpersonal continuity of care may not benefit patients in certain circumstances. Studies have found that interpersonal continuity did not improve completion of diabetes monitoring tests, 40 or lower HgbA1c values in diabetic patients. 41 Qualitative research performed with GPs in the United Kingdom found that the physicians felt obligated to maintain relationship with patients despite feeling powerless to improve clinical outcomes, even to the extent of colluding with unhealthy illness behaviours. 42 Further research is needed about negative consequences of continuity and patient’s perceptions of these outcomes in regards to safety.
One particularly interesting finding was that among individuals who had high trust in their physician’s deliberation, those who preferred deliberation in decision making were more likely to feel unsafe about seeing someone other than their regular physician. This finding suggests that patients feel safer when discussing treatment choices with an individual that they are used to seeing, and is consistent with other literature that suggests that a reason for interpersonal continuity preference is improved communication. 14 , 16
Our study has several limitations. As our sample population was predominantly white adults from a single geographical area with at least a high school education, this work may have limited generalizability to other populations. Our study population was predominantly clustered in the older adult range, which may have limited the ability for age to have statistical significance in this study. Future research is needed to explore characteristics of safety‐related interpersonal continuity preference in other groups. In addition, while safety is likely a multidimensional construct, we were limited to measuring safety in the unidimensional manner in which it was asked of participants in this study. Perceptions of safety are a complex issue worthy of further study. Nevertheless, our findings suggest that some patients link the doctor–patient relationship to perceptions of safety. Finally, our question asked patients how they would hypothetically feel about safety if interpersonal continuity were to be disrupted. We were unable to further explore how much of this perception was based on actual experience. Further research is needed about patients’ perceptions of safety with and without continuity. In particular, qualitative research might address when and under what circumstances individuals feels unsafe without continuity.
Our main findings—that some individuals perceive not being safe if they were to experience interpersonal continuity disruption, and that those individuals who have high trust in their doctor would feel unsafe about seeing someone else if they have preferences for deliberation about treatment choices—have important implications for the design of health‐care visits. When possible, office policies and procedures should be designed so that individuals who prefer interpersonal continuity should be scheduled with their own physician. Given the Institute of Medicine’s emphasis on shared physician–patient decision making, 43 , 44 disruptions to continuity in relationship should be minimized for those who prefer discussing treatment options with their own physician. Further research should identify reasons that patients feel unsafe about interpersonal continuity disruption and whether the perception of feeling unsafe is associated with subsequent negative health‐care outcomes.
No conflicts of interest exist for the authors.
This project was supported by the Community‐Academic Partnerships core of the University of Wisconsin Institute for Clinical and Translational Research (UW ICTR) funded through an NIH Clinical and Translational Science Award (CTSA), grant number 1 UL1 RR025011.
Portions of these results were presented on 10/16/06 at the 2006 North American Primary Care Research Group Annual Meeting in Tucson, Arizona.