Incidence and contributing factors to termination of the patient-physician relationship.

PURPOSE: Identify the incidence and factors contributing to the termination of gynecologic patient-physician relationships.
METHODS: All patients terminated from the practice between January 2008 and December 2012 were identified. Charts were reviewed for demographic information, termination reason, and cancer diagnosis.
RESULTS: In the five year study period, 8851 new patients presented to the division of gynecologic oncology. Within this cohort, 123 patient-physician relationships were terminated. Among terminated patients, missed appointments (63.4%), noncompliance to treatment (23.6%), disruptive behavior (10.6%), and drug abuse behavior (2.4%) were the key reasons for termination. While no patients were terminated for financial reason, statistical differences were found for those with Medicaid insurance (OR = 5; 95% CI: 3.4-7.1). Terminated patients were more likely to be younger, African American/Black, and have a diagnosis of GTD or cancer, particularly cervical cancer, when compared against all retained patients.
CONCLUSION: The prevalence of patient-physician relationship termination was low at 1.4% (123/8851). However, the finding that the 52% of terminated patients had a diagnosis of cancer is concerning; 73% of which had stage III or greater disease, or were unstaged. We hope that the identification and quantification of reasons for termination and those at risk for termination, as well as the introduction of patient-navigators, will lead to improved methods to ensure patient compliance and retention.

Introduction

From the moment an ailing person approached a knowledgeable advisor for aid, the patient-physician relationship has existed. Francis W. Peabody discusses the “vital importance of the personal relationship between physician and patient in the practice of medicine” in his landmark paper, The Care of the Patient (Peabody, 1927). He proposes that “the practice of medicine in its broadest sense includes the whole relationship of [the] physician with his [or her] patient”.

The American Medical Association (AMA) has sought to define this relationship and establish rules of conduct (American Medical Association, 1995). The focus is on the role of physicians to serve as their patients' advocate and foster patient rights. Six rights identified are the provision of information in a timely manner, allowing for autonomous decision-making, respectful treatment, preservation of confidentiality, ensuring continuity of care, and facilitating access to care. These principles of the patient-physician relationship become compromised when patients are noncompliant to treatment, follow-up, office policy, become verbally abusive, or unexplained nonpayment.

In the care of a patient's malignancy, as with other diseases, compliance with treatment and follow-up is of the utmost importance. Studies have demonstrated the physician-patient relationship to be closely associated with treatment adherence (Schneider et al., 2004). However, there is little data existing in the literature on contributing factors for termination of the patient-physician relationship. We sought to identify the incidence of patient termination and potential contributing factors in a gynecologic practice.

Methods

Terminated patients from January 2008 to December 2012 were identified and reviewed as part of a Gynecologic Oncology division quality improvement project. Study approval was obtained from our institution review board (BMHIRB1349) and determined exempt review. Terminated patients were those that were formally discharged from the division of Gynecologic Oncology. The process of termination was done in accordance with the American Medical Association's Code of Medical Ethics guidelines (Code of Medical Ethics, 1996). Patients were given written notice with certified mail and return receipt requested. The letter also expressed that treatment and access to services would be continued for 30 days in order to allow the patient time to secure care from another provider. The patient was given information and resources to help locate another alternative physician. Transfer of records to a newly-designated physician was also provided upon signed patient authorization to do so.

All patients during the study period were reviewed for demographic data including age, self-described race, and insurance status (uninsured, private, Medicare, Medicaid). For those patients that were terminated, a more in depth chart review was performed to obtain medical and social history (tobacco, alcohol, or drug use), reason for termination, and stage of malignancy, when applicable. Insurance status was defined as the patient's primary insurance. Tennessee Medicaid programs (Blue-care, TennCare, TLC, OmniCare, etc.) were grouped as Medicaid. Private insurances included Blue Cross Blue Shield, United Health Care, Humana, and other non-governmental insurances. Self-pay patients were classified as uninsured.

Primary diagnosis was defined as the medical diagnosis for which the patient was seen or referred to our clinic. New patients that were not terminated during the study period are referred to as retained patients; while terminated patients will be referred to as such. First documented visit to the last documented visit was used to define time to termination.

Disruptive behavior was defined as being rude, yelling, or threatening to staff and/or other patients. Missed appointments were defined as three or more consecutive no shows to clinic appointments despite phone calls and/or reminder letters. Noncompliance to treatment plan was defined as documented episodes of noncompliance to prescribed medications (including chemotherapy), not showing up to scheduled surgery, and failing to follow recommended treatment plan (including failure to show for radiation treatment appointments or surgery). Drug abuse included demanding excess pain medication, selling pain medications, or two or more positive urine tests for illegal drugs (cocaine, heroin, etc.).

Terminated and retained patients were compared to determine the significance of contributing factors. Two-sample t-test was used to compare continuous variables, and chi-square test was used to compare categorical variables. All analyses were done using SAS 9.3 (SAS Institute Inc., Cary, NC). Odds ratios were calculated to assess strength of association. Factors were considered to be statistically significant if p ≤ 0.05.

Results

From January 2008 to December 2012, the division of gynecologic oncology accepted 8851 new patients. Within this cohort 123 patient-physician relationships were terminated (Table 1). The median age of retained patients was 52 (range 11–102) compared to 47 years (range 19–94) in the terminated group (p < 0.001). The median time to termination was 144 days (1–1625). The most common reason for termination was missing appointments (78/123; 63.4%), followed by non-compliance to treatment, disruptive behavior, and drug abuse (Table 2). None of the terminations were for financial reasons. Race was statistically significant among the terminated patients (p < 0.001) with African American patients being more likely to be terminated (OR = 3.8; 95% CI: 2.6–5.5).

Table 1

Demographics of terminated and retained patients.

Factor	Terminated patients N (%)	Retained patients N (%)	OR (95% CI)
N	123	8728
Age (median, range)	47 (19–94)	52 (11 − 102)
Race
African American	80 (65)	2873 (32.9)	3.8 (2.6–5.5)
White	42 (34.1)	5363 (61.4)	0.33 (0.22–0.47)
Other	1 (0.8)	492 (5.6)	0.14 (0.01–0.69)
Insurance
Medicaid	50 (40.3)	1058 (12.1)	5 (3.4–7.1)
Medicare	35 (28.4)	2528 (29)	0.98 (0.5–1.4)
Private	29 (23.1)	4798 (55)	0.35 (0.16–0.38)
Uninsured	9 (8.2)	344 (3.9)	1.9 (0.91–3.7)
Primary gynecologic malignancy			2.2 (1.5–3.1)
No	59 (48)	5873 (67.3)
Yes	62 (52)	2855 (32.7)

Table 2

Features of terminated patients.

Characteristics	N	%
Terminated patients	123
Time of care (median days, range)	144 (1–1625)
Tobacco use	61	49.6
Drug abuse	19	15.4
Psychiatric diagnosis	33	26.9
Disability	19	15.4
Non-cancer diagnosis
Genital dysplasia	20	16.2
Pelvic/adnexal mass	16	9.7
Abnormal uterine bleeding	10	9
Endometrial hyperplasia	3	2.2
Fibroids/pain	2	0.7
Borderline ovarian tumour	2	1.5
Teratoma	1	0.7
Gestational trophoblastic disease	4	3.3
Reason for termination
Disruptive behavior	13	10.6
Drug abuse behavior	3	2.4
Missed appointments	78	63.4
Noncompliance to treatment	29	23.6

Those with a gynecologic malignancy were analyzed in Table 3. There were 62 gynecologic malignancies and two Krukenberg tumours. A diagnosis of cancer was a statistically significant risk factor for termination (OR = 2.2; 95% CI: 1.5–3.1), with cervical cancer and gestational trophoblastic disease (GTD) being the most significant. The for termination were 2.9 (95% CI: 1.8–4.8) and 8.1 (95% CI: 3.2–20.9), respectively. Seventy-three percent of terminated patients with cancer (45/64) were unstaged or stage III or greater. Of those terminated with cervical cancer, 70% were terminated due to missed appointments and 85% (23 / 27) had unstaged or stage IIB or greater disease.

Table 3

Cancer diagnosis and relative risk of termination.

Cancer site	Terminated patients N (%)	Retained patients N (%)	OR (95% CI)
Cervix	27 (22)	582 (6.7)	3 (1.8–4.9)
Ovary/fallopian/peritoneal	9 (7.3)	693 (7.9)	0.52 (0.24–1.0)
Uterus	21 (17.1)	1329 (15.2)	0.57 (0.33–0.97)
Vagina/vulva	5 (4.1)	216 (2.5)	1.1 (0.37–2.5)

Insurance status was a significant variable for termination (p < 0.001). Patients with Medicaid (OR = 5; 95% CI: 3.4–7.1) were more likely to be terminated; while private insurance was protective (OR = 0.35; 95% CI: 0.16–0.38). Among terminated Medicaid patients, 50% (15/50) had a diagnosis of cancer, with cervix being the most common site and 78% (39/50) were terminated for missed appointments.

Tobacco or alcohol use, drug abuse, psychiatric diagnosis, or being on disability was not associated with any particular reason for termination. Furthermore, patients with these characteristics made up a minority of terminated patients.

Discussion

We believe that improvement of a medical system begins with identification of a problem. For this reason, we undertook this study to identify the incidence of patient-physician relationship termination. The patient-physician relationship is the keystone of care that leads to optimum health outcomes (Dorr Goold & Lipkin, 1999). This study elucidates reasons for termination and some of the patient demographics within the terminated group. Younger age, African American race, Medicaid insurance, GTD, and a diagnosis of cancer, particularly cervical cancer, were significantly associated with termination compared to retained patients. It is concerning that 64 of the 123 terminated patients had a diagnosis of cancer, 25 of which had stage III or greater disease and 20 were not staged. Among cancer patients, the overwhelming reasons for termination were missed appointments and noncompliance to treatment (91%). We postulate that socio-economic status may be an underlying cause. In the case of missed appointments, several factors may be at play. A lack of resources may limit a patient's access to transportation, ability to miss work, or find childcare in order to keep appointments. This may also reflect a poor familial and/or social support system. It is interesting to note that our non-paying patients, whom we presumed to be financially poor, did not have significant compliance issues. Other social issues (disruptive behavior, drug abuse, etc.) may contribute to patient termination more than just poverty alone.

The four main factors for dismissal found in this study (disruptive behavior, drug abuse behavior, missed appointments, and non-compliance to treatment) serve as launching points to begin addressing the underlying problem and to begin to address the dispositions participants bring from experience to the clinical encounter.

The main weakness of our study is the retrospective nature of the study. We were not able to question the terminated patients on the reasons attributed to dismissal or to assess the patient's perception thereof. Once the termination letter was sent, our group recommends physicians have no further contact with the terminated patient, except for extenuating circumstances after termination (i.e. the physician receives records and needs to direct them to the patient, the patient was on a medication that now has been recalled and the physician needs to contact the patient for further instructions, or for release of records). As a result, we are unable to assess follow-up of these patients.

From internal discussion among our physicians and nurses, the complexity of medical care and the multidisciplinary approach often used in cancer care may be difficult for certain patients. Patient-navigator assistant program may be able to aid patients. We hope these assistants will be able to establish a personal relationship with patients and facilitate the resolution of social, economic, and personal issues that may otherwise limit care. We plan to repeat this study in the future to evaluate if this program is successful in reducing our termination rate further.

Our study identified the incidence of patient-physician relationship termination and revealed potential factors that may contribute to termination: age, race, type of insurance, and a diagnosis of malignancy, particularly cervical cancer.