Continuity and Trust in Primary Care: A Qualitative Study Informed by Game Theory

Tarrant and colleagues[1] are to be commended for taking the slippery issue of trust and applying a theoretical model to better understand its relationship with patient-doctor continuity. Continuity of patient care is something that continues to define family practice internationally,[2] but evidence that it leads to improved patient outcomes is weak.[3] The literature emphasises the importance of interpersonal aspects of longitudinal patient-doctor relationships in determining the value of longitudinal care to patients and doctors,[4] yet personal continuity has mainly been measured in terms of the number or proportions of consultations with the same doctor.[5]

Measures of trust have been developed,[6] which may represent the value of on-going relationships better, but as this article highlights, there is also the issue of personal knowledge. Whilst a doctor may see a new patient with the benefit of “institutional trust” (building up to “secure trust” over repeated encounters), at first encounter patient and doctor both start with a much more limited knowledge base: the doctor may have the medical information contained in the patient’s records and the patient may know something of the physician from indirect encounters or “conversations over the fence” with their neighbour.[7] Doctors and patients emphasise the value of knowing patient’s personality, family and social circumstances[8] – information which is usually only gleaned from personal contact with the patient over time. In addition, this paper flags up the issue of doctor trust of the patient, but obviously does not attempt to apply game theory from the physician’s perspective. Therefore, further studies might want to examine trust from the other side of the relationship, and “true” measures of personal continuity (depth of relationship) might need to incorporate personal knowledge as well as trust.[7]

(1) Tarrant C, xon-Woods M, Colman AM, Stokes T. Continuity and Trust in Primary Care: A Qualitative Study Informed by Game Theory. Annals of Family Medicine. 2010;8:440-446.
(2) Allen J, Gay B, Crebolder H, Heyrman J, Svab I, Ram P. The European definition of general practice/family medicine. Barcelona: WONCA Europe; 2002.
(3) Saultz JW, Lochner J. Interpersonal continuity of care and care outcomes: a critical review. Ann Fam Med. 2005;3:159-166.
(4) Freeman G, Hjortdahl P. What future for continuity of care in general practice? BMJ. 1997;314:1870.
(5) Saultz JW. Defining and measuring interpersonal continuity of care. Ann Fam Med. 2003;1:134-143.
(6) Calnan M, Rowe R. Trust in health care: an agenda for future research. London: The Nuffield Trust; 2004.
(7) Ridd M, Shaw A, Lewis G, Salisbury C. The patient-doctor relationship: a synthesis of the qualitative literature on patients' perspectives. BJGP. 2009;59:268-275.
(8) Tarrant C, Windridge K, Boulton M, Baker R, Freeman G. Qualitative study of the meaning of personal care in general practice. BMJ. 2003;326:1310.

Competing interests:   None declared

Tarrant and colleagues make several excellent points re the relationship between repeated interactions and secure trust. I would like to expand on the implications of their work and offer additional support for their findings from work I have done with colleagues over the past 15 years.

It has been noted that the association between continuity, particularly the length of the relationship, and level of trust could be the result of continuity leading to greater trust, or could represent the effects of patients with lower initial levels of trust leaving the relationship (Thom et al, 1999). The study by Tarrant et al provides evidence that continuity, by allowing for repeated interactions, can increase trust. While repeated interactions per se seem to have value in fostering trust, patients in our focus groups sometimes recalled a key, ‘high stakes’ interactions that demonstrated their doctor’s competence and caring that significantly raised their level of trust (for example an urgent referral to cardiology for unstable angina).

Tarrant and colleagues also mention the value of continuity in enabling patients to “build their own reciprocal reputation with the GP.” The role of mutual trust has been investigated in the social science literature but not, to my knowledge, in medicine. In interviews and focus groups we have conducted, physicians identified mutual trust as an important aspect of the relationship quality with potential effects on health care outcomes (Stepanakova et al, 2009). Further investigation of the role of mutual or reciprocal trust would help characterize the nature of this important dimension of the patient-physician relationship.

As Tarrant et al. point out, game theory research provides important insights that can help guide future research on patient trust. For example, investigators have identified several ‘exchange’ characteristics that promote interpersonal trust in the experimental setting (Deutsch 1962, Gambetta 1988) that seem relevant to the doctor-patient relationship: greater perceived mutual interests; opportunities for communication; less risk for unfulfilled trust; a previous history of fulfilled trust; less perceived difference in power; and an expectation of a longer term relationship (Thom 2000). These factors can be measured and potentially modified through clinician education and training. Hopefully the article by Tarrant and colleagues will promote further work of this type.

References:
Deutsch M. Cooperation and trust: some theoretical notes. Nebraska Symposium on Motivation, 1962, pp 275-319.
Gambetta D. Trust: Making and Breaking Cooperative Relations. Cambridge, MA. Oxford Press, 1988.
Stepanikova I, Cook KS, Thom DH, Kramer RM, Mollborn SB. Trust in Managed Care Settings: Physicians’ Perspective. In: Who Can We Trust? How Groups, Networks, and Institutions Make Trust Possible. Cook KS, Levi M, Hardin R (Eds). Rusell Sage Foundation, Thousand Oaks, CA, 2009.
Thom DH, Ribisl KM, Stewart AL, Luke DA. Validation of a measure of patients’ trust in their physician: the Trust in Physician Scale. Med Care 1999;37:510-7.
Thom DH. Training physicians to increase patient trust. J Eval Clinical Pract 2000;6:249-55.

Competing interests:   None declared

I read Tarrant, Dixon-Woods, Colman, and Stokes’ article with great interest. I applaud their creative use of game theory to illuminate aspects of trust development in physician-patient relationships. Game theory provides a useful yet surprisingly neglected framework for thinking about these issues. In particular, explicating the role repeated interactions play in the development of what Tarrant et al term secure trust is enormously important. This study will hopefully stimulate a great deal of follow-up research.

The game theory concept of repeated interaction, we should note, was developed in the context of theory and experimental studies involving simple (and usually short) binary choice games, where the stark decision is between cooperating or trusting fully--or not at all. In the context of ongoing physician-patient relationships, the choices are much more complex, nuanced, and drawn out. Accordingly, the empirical question arises, “What precisely are the elements of repeated exchange between physician and patient that are most likely to facilitate the development of secure trust?” Social psychologists have investigated this question in the context of what they term history-based trust (Kramer, 1999). What does that research suggest? An impressive body of work indicates that demonstrations of concern and empathy for the other person in the relationship matters considerably. Demonstrations of respect for the other, affirming the dignity and status of the other, and providing opportunities for the other to have voice in the relationship have also been found to be very important in trust-building.

Although these may be important factors, they are also quite subtle things to communicate. This raises the question, “What can physicians do to communicate concern, empathy, and respect for their patients?” In our research on the determinants of perceived trustworthiness and trust development in physician-patient relationships, we found strong evidence that effective verbal and non-verbal behaviors for increasing trust included such simple matters as direct eye contact, allowing adequate time for patients to both ask questions and answer questions thoroughly, reducing the physical distance between the patient and the physician during their interactions, and the use of supportive physical contact (e.g., touching a patient’s arm in order to show concern or interest) (Cook 2004).

These verbal and non-verbal behaviors are important also because research suggests that individuals often are not very conscious of them—and especially the magnitude of their impact on others. Our research, and that of others, suggests that trust building behaviors are learnable, just like any other vital skill in the physician’s tool-kit.

References
Cook, K. Kramer, R. M., & Thom, D. (2004). Trust and distrust in physician-patient relationships: Antecedents and consequences. In R. M. Kramer & K. S. Cook (Eds.), Trust in Organizations: Dilemmas and Approaches. (Russell Sage Foundation Trust Series, Volume VII) New York: Russell Sage Foundation.

Kramer, R. M. (1999). Trust and distrust in organizations: Emerging perspectives, enduring questions. Annual Review of Psychology, 50, 569- 598.

Competing interests:   None declared

I am delighted to see this paper. As the authors say on p 441, ‘A key problem is that empirical research into continuity of care has lacked an integrative theory’. I do agree with them that ‘game theory addresses this lacuna’! Governmental pressure in favour of institutional trust at the expense of trust in an individual clinician is widely prevalent. I have argued before in your pages that ‘quantitative measures merely tell us about contacts, they say nothing about whether the relationship (if any) was therapeutic’ (1). Likewise here (p445) we read ‘Our findings suggest that repeated interactions over time are not in themselves sufficient to develop secure trust. Though some degree of longitudinality is needed, the development of trust is also dependent on what happens in interactions.’ This is challenging because it forces us to assess the content of clinical transactions and not just record that they’ve happened.

Here in London we are currently working with negotiators and policymakers to try and introduce worthwhile incentives for Family Practitioners/General Practitioners to prioritise continuity of care, rather than merely quick access to any available clinician. Game theory as described here enables us to argue our case more strongly! Together with agency theory, so persuasively explained by Donaldson (2), it also provides a firm basis for future research into how we can improve trust between patients and clinicians. This is a desirable outcome on the way to health gain and not ‘just a process’.

1 Freeman GK Importance of measuring the interpersonal element of continuity - commenting on Saultz JW, Albedawi W. Interpersonal continuity of care and patient satisfaction: a critical review. Ann Fam Med. 2004;2:445-451. http://www.annfammed.org/cgi/eletters/2/5/445 9th Nov 2004.
2 Donaldson MS. Continuity of Care: A Reconceptualization. Medical Care Research and Review 2001;58:255-90

Competing interests:   I have worked with the first author on continuity of care research for the last 10 years

Tarrant et al. (2010) consider the effect of changes in the UK primary care system on the trust patients place in their general practitioners. Tarrant et al. report the results obtained from a series of interviews with patients in regards to the trust they have in their GPs, contrasting trust under single patient-physician interactions with trust under repeated patient-physician interactions. They conclude that in a single interaction system trust is drawn primarily from institutions (i.e. they trust that the system will ensure physicians are competent and act in their best interest) and in repeated interactions trust may be enhanced by reputation building on the part of the physician. Tarrant et al. claim that these findings are consistent with the predictions of game theory.

While game theory offers a powerful tool for policy analysis, its use and reference in Tarrant et al. is misleading. Game theory is a way of modeling and thinking about the world, not a single theory with general conclusions. The notion of trust that is studied in game theory regards trust in behavior, not trust in ability, as is relevant to Tarrant et al. The scenario Tarrant et al. examine can be understood in a simpler asymmetric information framework, in which patients are uncertain about physician abilities. By invoking game theory Tarrant et al. are implicitly assuming that physicians have an incentive to not serve their patients best interests; an assumption that is not supported by their empirical evidence. This paper provides insufficient information to determine if the findings referenced from game theory studies apply in this context.

The authors provide a broad and superficial description of game theory that overstates the generality of their argument. Game theory does not yield many universal results, as outcomes are dependent on the structure of the game under consideration. Statements such as “… as game theory would predict, a history of past interactions and the anticipation of future interactions provide a context that makes it possible for trust to build and become secure” (p.445) refer to the result of a specific game (or a family of games), not to game theory in general.

It is not possible to determine which game, or family of games, the authors are considering since they fail to sufficiently describe the patient-physician “game”. Game theory requires four basic elements that must be described in any application:[i] There must be a set of players; a set of rules describing the actions and information available to each player; a set of outcomes stating what happens for any combination of actions; and a set of payoffs for each player at any possible outcome. Given these elements, one can determine player strategies and make predictions about the outcomes expected to occur. As Tarrant et al. do not describe how these basic elements map to their context, there is no evidence that the situation examined in this study corresponds to a game.

Talking about trust is not equivalent to talking about game theory. In game theory, trust is used to explain the results of a subset of games but game theory does not explain the origins of trust. Statements such as “…the game theory suggestion that there is little inherent in a single interaction to provide a secure foundation for trust.” (p.443) are inaccurate. In game theory, “trust” is contingent on the payoff to each player—one player must trust another to act in their best interest even though the other player has the option, if not the incentive, to do otherwise. In the absence of specified strategies of physicians and pay-offs to those strategies, “trust” in Tarrant et al. seems to be more about a patient’s trust in the competence of the physician rather than a patient trusting a doctor to act in the patient’s best interest. This is an information problem, not a game theoretic problem.[ii] There is no game because only one player (the patient) is making any moves (visit or not visit). Trust compensates for lack of information; repeated visits change the information, but not necessarily the trust.

i. For an introduction to games we recommend Dixit, A., Skeath,S., Reiley, D., Games of Strategy. Norton, 3rd Ed., 2009.
ii. This type of information problem is considered in Akerlof G.A., 1970. The Market for “Lemons”: Quality Uncertainty and the Market Mechanism. Quarterly Journal of Economics 84(4), 488–400.

Jesse Matheson and Daniel J. Dutton, Department of Community Health Sciences, Faculty of Medicine, University of Calgary

Competing interests:   None declared