- Email: [email protected]
Practices, strategies, and motivations in treatment of rheumatoid arthritis
Practices, Strategies, and Motivations in Treatment of Rheumatoid Arthritis
ALFRED E. GOLDMAN, Ph.D. SUSAN SCHWARTZ MCDONALD, Ph.D. Philadelphia, Pennsylvania
From National Analysts, Division of &x-Allen & Hamilton, Incorporated, Philadelphia, Pennsylvania. Requests for reprints should be addressed to Dr. Alfred E. Goldman, National Analysts. 400 Market Street, Philadelphia, Pennsylvania 19106.
National Analysts conducted primary research with rheumatoiogists-specificaiiy, two panel discussions, 25 in-depth telephone interviews, and a mail survey of conference (Auranofin Symposium and Workshop) participants-to examine current treatment practices and to probe the rationale and motivations underlying treatment strategies in rheumatoid arthritis. The research identified important areas of consensus in drug perceptions, therapeutic approaches, and disagreements. Physicians differ regarding the mlnfmum time they wait after diagnosing rheumatoid arthritis before initiating remittive therapy, some beginning immediately and others waiting six months or longer. Younger physicians are quicker to initiate remittive treatment than their older colleagues, but both younger and older practitioners are initiating remittive therapy earlier than in the past. Some noteworthy differences between hoepitai-based and office-based practitioners were discerned with resp&t to factors that figure in their decisions to initiate remittive therapy. Differences were also found among physicians in the way they pose drug options to their patients; “authoritarian,” “iibertarian,” and “guided democracy” were names given to the three styles ktentMd. in general, however, physicians report that patients are more directly involved in treatment selection than previously, a trend that may in part be due to the use of more aggressive treatment strategies than in the past and a desire to share the psychoiogk burden of those decisions. Findings suggest that goid w will continue to be a mainstay first-line disease-modifying agent in the treatment of rheumatoid arthritis but that there may be less reluctance to use other agents as physicians become increasingly familiar and comfortable with aiternative options, especially penkiiiamine and immunosuppressive agents. At the request of the Auranofin Symposium executive committee, National Analysts, a marketing research and consulting organization, conducted research on the way clinicians confront rheumatoid arthritis, with the express purpose of sharing the results with the conference participants. The data collected provide a descriptive picture of how practitioners are dealing with rheumatoid arthritis, how they view the available treatment options, and the basis for their decisions in treating individual patients. The mix of techniques selected-“quantitative” procedures (which lend themselves to statistical analyses) and “qualitative” procedures (which address attitude structure rather than frequency)-together provide a measure of underlying motivation as well as a count of clinical practices.
December 30, 1983
The American Journal of Medlclne
79
ORAL GOLD SYMPOSIUM-GOLDMAN
TABLE
I
and MCDONALD
What is Minimum Length of Time You Would Prefer to Wait After a Confirmed Diagnosis of Rheumatoid Arthritis Has Been Made, Before Initiating Ftemittive Therapy? Percent
Reply
No minimum-stat-t
immediately
if warranted
54 24 16 2 2
3 months 6 months 12 months More than 12 months
METHODS
The findings reported here reflect three complementary methods: (1) two group discussions with rheumatologists; (2) approximately 25 in-depth telephone interviews with rheumatologists across the country: and (3) a mail survey of 121 conference participants, 94 percent of whom are rheumatologists. Group Discussions. Group interviews involved eight to 10 practitioners assembled in the presence of a trained moderator. The sessions lasted about two hours, capitalizing on the principles of group dynamics and peer interaction to elicit responses of a depth that would not be forthcoming in a survey. Valued for their sensitivity and candor, group sessions do not provide statistically projectable data; they fall instead under the heading “qualitative research.” Two sessions were conducted, one in Philadelphia and the other in Atlanta. Like group sessions, indiIndividual In-Depth Interviews. vidual telephone interviews are qualitaltive in nature and are not appropriate for statistical projection. They do, however, permit a more exhaustive and intensive conversation and broader geographic representation.
% Who Would Start Remittive Therapy Immediately, If Warranted 75% (61%)
50%
.b)
25%
0% Under 40
40 - 49 PHYSICIAN
50 + AGE
Figure 1. Younger physicians, as a group, tend to initiate remitting drug therapy sooner than their older colleagues.
90
December 30, 1993
The American Journal of Medicine
Mail Survey. All those who accepted the invitationto attend the symposium were sent questionnaires and asked to participate in the survey. Of the nearly 130 who did attend the sessions, 121 (or better than 90 percent) returned questionnaires in time to be processed prior to the symposium. Thus, responses reported here represent a view of clinical practices before exposure to conference material and constitute a virtual “census” of the audience that was present. Because only a few respondents in the survey sample did not describe themselves as rheumatologists, we were unable to compare the responses of different specialty groups. We did, however, make use of another practice variable-office-based versus hospital-based-a classification scheme reflecting the way physicians defined themselves when given those options in the questionnaire. For purposes of analysis, the responses of solo and group (office-based) practitioners were combined. The term “remittive therapy,” although not specifically defined, was equated in the questionnaire with the following: injectable gold, hydroxychioroquine, peniciilamine, and im-
munosuppressive agents. RESULTS Along with some areas of broad consensus, the data suggest noteworthy differences in treatment styles and strategies-variations that seem to reflect the different ways physicians have of resolving the competing values and perils endemic to treatment. Both the group and individual in-depth interviews suggested that rheumatologists share a sense of frustration and vulnerability in treating rheumatoid arthritis, a disease described by one practitioner as “monstrous.” The best they could hope for is to halt or delay progression of the disease with drugs that may or may not work and that pose alarming risks. There was disagreement about when to initiate therapy with a remittive agent, which surfaced in the survey as well as in the qualitative phase. Some begin therapy promptly, as soon as they suspect that a patient has rheumatoid arthritis, because they believe that to wait for evidence of bone erosion is, as one put it, “to lose the game.” Others prefer to wait, sometimes for months, before initiating remittive therapy. As Table I illustrates, 54 percent of the physicians surveyed would start therapy “immediately if warranted” once a confirmed diagnosis had been made, 24 percent would wait a minimum of three months, and another 18 percent would wait six months. Only 4 percent would wait a year or longer. Personal discussions identified at least one important reason for the variation, the possibility of spontaneous remission. Some prefer to temporarily defer treatment, hoping the disease will halt independently. Factors like physician age, recency of training, and institution also seem to affect the speed with which they begin remittive treatment. Younger physicians were quicker to initiate therapy than their older colleagues: 61 percent
and MCDONALD
ORAL GOLD SYMPOSIUM-GOLDMAN
of those under age 40 would initiate therapy immediately, whereas only 53 percent of those in their 40s and 48 percent of those over 40 would make the same decision (Figure 1). Roughly half the respondents surveyed, younger and older practitioners alike, claimed to be initiating remittive therapy earlier than they used to, suggesting that the perils of the disease are outweighing the perils of the treatment in these physicians’ clinical equation. No one is waiting longer to initiate remittive therapy than in the past. The primary factors precipitating a decision to begin remittive therapy are of interest. Physicians were virtually unanimous in emphasizing bone erosion, failure to respond to nonsteroidal anti-inflammatory drugs, and joint swelling as triggers in that decision, but there were some striking differences in the priorities of officebased and hospital-based practitioners (Table II). Office-based physicians cited a wider number of factors, including nonphysiologic ones, as “very important” in this decision to begin remittive therapy. Specifically, more office-based practitioners considered threatened job loss, frequency of flare-ups, curtailment of activities,
50%
1
TABLE II
Mean Importance of Top 10 Factors in Decision to Initiate Remittive Therapy Mean’
Factor Increasing number and severity of swollen joints Not responsive to nonsteroidal anti-inflammatory drugs Presence of bone erosion More frequent flare-ups Presence of other extra-articular manifestations Curtailment of activities Threatened job loss Unrelenting pain Increased fatigue, malaise, weakness Joint space narrowing All others
5.3 5.2 5.2 4.6 4.6 4.5 4.5 4.4 4.2 4.2 <4.0
’ On scale from 1 to 6, where 1 = not at all important and 6 = very important.
and unrelenting pain to be “very important” in making their decision. Figure 2 lists those factors and the percentage who deemed them “very important,” using a 6-point scale. We can only speculate on the reasons for this difference: since hospital-based physicians are, on
THREATENED JOB LOSS
CURTAILMENT OF ACTIVITIES 50%
40
.., . .,..... ::::: L I
30
(25%)
1
20
10
0
Office Based
Hospital Eased
::::::::::::::::::: ::::::::::::::::::: ......,...,....,. .::::::.:::::::::::: .....).
:::::>..:.:.:.:.:.: ::::::y::::::::: (10%) :::y$:::::::y: :::>.::::y::$;>, :::y:::::::::::> ::::;::::::::::::: ::::::::::::::.:.: ......#.......L..
Office Based
UNRELENTING PAIN 50%
Hospital Based
MORE FREQUENT FLARE UPS
50%
1
:>.~.~.~.~.>.~.~. ::::::::::::::::::: >:s’c.:.:.:.:.>: :y::::::::::::::: .‘.....S........‘.
1
40.
30 -
30
(27%)
-
20
Figure 2. Importance of factors in decision to initiate remittive therapy for office- and hospital-based physicians. (Percent citing factor as ‘very important, ’ or ‘6’ on a 6 point scale.)
10
0 b
-
Office Eased
Hospital Eased
December 30, 1983
(22%) .. ::::::::::::j:::: ::::::::;::::::::: ..::::_.:.~.. ..........~...~.~.~.~.. ......... :::::::::::::::::: .. . (10%) :.:.:.:.:.:.:.:.:.
. . . :.:.:.;.:.:.:.:.:. :::::::::::::::::: :::::::::::::::::: .:.:.:.:.:.>:.:.: ;:>i:):$:>;:;: Office Based
... ..,.. :::::::::::y:::: .:+>:.:.:.:.:.: ::y::::$.:.:. :...>:.:.::::..:..: .>:.>:.:.:.:.:.:
Hospital Based
The American Journal of Medicine
81
ORAL GOLD SYMPOSIUM-GOLDMAN and MCDONALD
TABLE Ill
Characteristics Associated with Key Remittlve Agents (Percent Assigning Description to Each Drug) Percent
Characteristic
Injectable Gold Most Most Most Most
effective generally effective for aggressive disease effective for smouldering disease preferable for young adults
90 a4 79 78
Hydroxychloroquine Least costly Causes least frequent side effects Least effective generally Causes fewest serious side effects
96 96 90
88
Penicillamine Most difficult to tolerate over long period Causes most frequent side effects Causes most serious side effects
average, slightly less aggressive
80
78 76
in initiating remittive
therapy, they may, in fact, be less responsive to what they believe are more “peripheral” considerations. It is also possible that office-based physicians know their patients better and that the intimacy or continuity of that relationship makes patients’ physical discomfort and psychologic stress seem more urgent. There was general agreement about key characteristics of first-line remittive agents, particularly such
features as efficacy, toxicity, and safety. The vast majority designated injectable gold as “most effective generally” for either aggressive or smouldering disease, most preferable for young adults, and most costly (Table Ill). At least eight in 10 termed hydroxychloroquine most convenient and safest (causing least frequent and fewest serious side effects) but also least generally effective. Finally, most viewed penicillamine as least safe and most difficult to tolerate over a long period of time. Consistent with those drug profiles, most of the survey respondents initiate remittive therapy with gold in the majority of patients. Table IV reports the average percentage of patients in whom gold, hydroxychloro-
TABLE IV
Initiating Drugs for Remittive Therapy
Percent of MDs ever starting patients on Injectable gold Hydroxychloroquine Penicillamine Immunosuppressive
agents
100% 88 % 76% 27%
Average percentage of patients receiving remittive therapy who were started on Injectablegold Hydroxychloroquine Penicillamine Immunosuppressive agents
71% 16% 14% 2%
82
December 30, 1883
The American Journal of Medicine
quine, and penicillamine therapy was started. However, some important variations in treatment style and drug perspectives emerged on further examination of the figures. Because physicians perceive treatment with remittive agents to be fraught with idiosyncracy and risk, and because patients often have strong feelings about the mode of therapy, a rigid progression of drugs or fixed treatment hierarchy could not be identified. Physicians who participated in the qualitative phase described a number of relevant considerations. The disease stage and rate of progression, the patient’s age and cardiovascular condition, and his or her mental set (especially reliability and anticipated compliance) were all factors that militated against one agent and for another. For example, some patients were more fearful of gold injections than of the serious, even life-threatening, side effects that oral alternatives might precipitate. That fear could be decisive in eliminating gold therapy as a serious option. By contrast, some patients came to their doctors expecting gold shots and that expectation guided the clinician’s course of action. Some physician variables also were identified. In particular, three distinct physician orientations to treatment were encountered in the qualitative analysis. One approach, labeled “authoritarian,” offers a decisive course of treatment, saying to the patient, “This is the drug I want you to take, and these are my reasons.” The second, more “libertarian” approach offers the patient a menu of acceptable options; it points out the strengths and the liabilities of each one and then says, “Now you may choose.” The third is perhaps best described as “guided democratic.” This approach is similar to the libertarian; it offers the patient a series of options, outlines the advantages and risks attached to each, but then communicates more or less directly the physician’s preference. Like many classification schemes, this one should be viewed as a continuum rather than a set of rigid categories. The authoritarian model is apt to yield under patient pressure, and the libertarian model may be a rigged game in which the physician successfully imposes his preference, however subtly. We believe that these differences can have important implications for therapy. Indeed, the survey of symposium participants provided some data to substantiate or quantify the typology. The majority (64 percent) reported that they “tend to select a single drug and present that to the patient,” whereas the rest indicated that they “tend to present the patient with two or more options.” The less structured (qualitative) discussions with physicians suggest that the tendency to involve patients more directly and explicitly in the decision-making process is growing.
ORAL GOLD SYMPOSIUM-GOLDMAN
This trend may be in part a reflection of more aggressive use of remittive agents and first hand familiarity with a wider array of options. It is consistent with movement toward more fastidious informed consent in treatment protocols. More importantly, it is a mechanism for sharing some of the burden in the decisionmaking process. Moreover, as many physicians pointed out in the in-depth interviews, rheumatoid arthritis therapy is really a lifetime collaborative effort between patient and physician; without the wholehearted commitment and compliance of the patient, success of the treatment is seriously compromised. Involvement of the patient in choosing among therapeutic options may be an attempt to enlist the patient’s commitment to the elected course of treatment. Practitioners differ from one another in other important ways. Although all are privy to the epidemiologic data about incidence and severity of drug side effects, they come to varying conclusions about the risks actually posed to their own patients and about the degree to which those risks are tolerable. The differences reflect variations in the way physicians weigh statistics versus clinical perceptions-in other words, how they assess “actual” risk in relation to “perceived” vulnerability. A classic illustration of the difference between statistical and perceived risk is the fear occasioned in some people by airplane travel, even though they recognize that, statistically speaking, the car trip to the airport is more dangerous. In a similar vein, nearly all the physicians surveyed agree that hydroxychloroquine is generally the safest drug, yet some prefer to use penicillamine because the small risk of permanent blindness from hydroxychloroquine is unacceptable either to them or to their patients, especially given what some physicians view as the questionable efficacy of that drug. Others would prefer to play these very complex odds another way; with careful monitoring, they are entirely comfortable in the belief that major problems with hydroxychloroquine can be detected and reversed. Still others (27 percent of the sample) initiated remittive treatment of a select handful of patients (usually fewer than 5 percent) with immunosuppressive agents, whereas the rest of our survey respondents rejected that level of risk (Table IV). Clinical experience, residency training, institutional CUStOfW personality, patient demographics, and a host of related factors influence those assessments; as physicians made clear in every phase of our research, every new patient invites some recalculation of the odds. In fact, besides providing information on physicians’ treatment strategies, the survey also provided an opportunity to compare approaches in the management of several “case studies.”
MANAGEMENT
and MCDONALD
PROBLEMS
Case 1. Physicians were asked how they would respond to evidence of an increase in bone erosion and cartilage loss in a 45year-old woman with moderately aggressive rheumatoid arthritis who was experiencing intermittent flare-ups affecting mostly the hands, wrists, and ankles. All but one of the respondents were prepared to treat this patient with remittive agents and, of those, 94 percent recommended injectable gold. Nearly half-approximately 46 percent-would offer the patient several options. Almost four in 10 would offer penicillamine, and more than a quarter would also recommend hydroxychloroquine. Two were prepared to initiate immunosuppressive therapy. (Note that in Figures 3 through 5, which illustrate the case study responses graphically, the encircled drugs are those that would be offered to the patient along with the percentage who would recommend each drug. More than one drug could be selected if physicians would choose to present several options. The modal daily starting dose is linked to each drug by an arrow.) There was a fair degree of consensus in the starting dosages for any given drug: nine in 10 would give 50 mg or more per week of gold, six in 10 would recommend 400 mg of hydroxychloroquinine (except for injectable gold, all drug dosages cited are per diem.), and seven in 10 would begin with 250 mg of penicillamine. The remainder would initiate therapy at more conservative dosage levels. This represents a fair degree of consensus because, from a statistical point of view, large majorities indicated that they would use the same dosage. On the other hand, there was a sizeable minority in disagreement. Indeed, our group discussions
88% 50mg.
Figure 3. Treatment strategies in a 45-year-old woman (Case 1) with moderately aggressive rheumatoid arthritis, intermittent flare-ups, increase in bone erosion and cartilage loss.
December 30, 1983
The American Journal of Mediclne
83
ORAL
GOLD
SYMPOSIUM-GOLDMAN
and MCDONALD
n
PBNICIUJMINE
MAXIMUM
F@re 4.
Treatment strategies on an engineer (Case 2) with
mild nagging joint symptoms and significant ear/y morning stiffness; currently receiving prednisone (5 mg per day) and
aspirin.
J
Figure 5, Treatment strategies for a 33-year-old woman (Case 3) with aggressive rheumatoid arthritis: no improvement with gold, visual symptoms with hydroxychloroquine.
INJECTABLE GOLD
HYDROXYCHLOROQUJNE
% Rspa-ting
Increased UseOf:
PENICILLAMINE
IMMUNOSWPRESSIVES
75% -
75% 7 Ki%I
50%.
PENICILLAMINE
IMMUNOSUPPRESSIVES
Undw40
40.49
PHYSICIAN AGE
1
50+
Under 40
40.49
50+
PHYSICIAN AGE
Figure 6.
Percentage of physicians reporting increased use of remittive agents.
F&e 7. Relationship between physician age and changing drug usage. Older physicians are using more penicillamine and fewer immunosuppressive agents than younger ones.
suggest that practitioners treating rheumatoid arthritis often assume greater consensus than really exists. Case 2. There was considerably less agreement as to whether a remittive agent was appropriate for the second patient, an electrical engineer with mild, nagging joint symptoms and significant early morning stiffness. One fifth of the physicians indicated that this patient was not yet a candidate for remittive therapy. Those who did prescribe a remittive agent were most likely to recommend gold (84 percent) or hydroxychioroquine (47 percent). Only about 20 percent thought that peniciliamine was warranted. There was consensus on the appropriate starting dose for gold (88 percent would
recommend 50 mg or more) but less agreement on dosage strategy for hydroxychioroquine. Case 3. The third case drew near unanimity: 91 percent were prepared to recommend peniciiiamine to this 33-year-old woman with aggressive rheumatoid arthritis, who had not experienced any improvement with gold therapy and had had visual symptoms with hydroxychioroquine. Another 44 percent were prepared to recommend immunosuppressive agents, with preference divided roughly evenly between azathioprine and methotrexate. The modal daily starting dose for azathioprine was 50 mg, and approximately 1 mg for methotrexate. (Note: Many physicians appear to be
84
December 30.1383
The American Journal of Medlclne
ORAL GOLD SYMPOSIUM-GOLDMAN
and MCDONALD
splitting methotrexate dosages and/or prescribing an alternate-day or a weekly regimen.) Many, however, were willing to exceed 100 mg of azathioprine (37 percent) or 3 mg of methotrexate (40 percent) per day (Figure 5). These numbers offer a synopsis of current practices, and a possible forecast of future treatment strategies. There is some evidence to suggest that physicians are increasingly receptive to the use of immunosuppressive agents. They are also using more penicillamine than in the past. Figure 6 indicates percentages of the survey sample reporting either less frequent or more frequent use of all major remittive agents; gold and hydroxychloroquine are, more or less, holding their own. Of particular interest is the fact that younger physicians are most apt to increase their use of immunosuppressive agents, whereas older physicians are reporting growing use of penicillamine (Figure 7). Those figures forecast the near-term, and perhaps long-term, trend in the treatment of rheumatoid arthritis: physicians over 40, having had substantial experience in their practices with penicillamine, are becoming less leery of that drug; their colleagues under 40, fresh from closer contact with immunosuppressive agents in their training, are discounting or accepting some of the reputed dangers of those agents.
similar data against which to compare these findings exist, and there is no way of judging what the responses to these questions might have been several years earlier. These data represent a “snapshot” taken at one point in time depicting a selected universe. It is tempting, however, to extrapolate and suggest that remittive agents like penicillamine and immunosuppressives will be used increasingly as physicians take a more aggressive posture in treating rheumatoid arthritis. Admittedly, the younger physicians who indicate greater willingness to prescribe immunosuppressive agents may, with age, grow to act more like their older, more cautious colleagues. On the other hand, familiarity may breed some comfort, if not contentment, with the second- and third-tier drugs, provided nothing occurs to dramatically alter the risk scenario or obviate the need for these agents. If there is one conclusion that can be safely drawn, it is a familiar one: reasonable practitioners will continue to differ in the treatment strategies they devise for rheumatoid arthritis. Those differences are presently bound by shared constraints and anxieties. Future treatment for rheumatoid arthritis is likely to remain a careful, vigilant course between the perils of the disease and the continuing perils of treatment.
COMMENTS
ACKNOWLEDGMENT
One of the misfortunes of medical survey research is that it is seldom published where physicians will encounter it because it is not designed to educate physicians about their own behavior. To our knowledge, no
We would like to thank Maria Figueroa of National Analysts for her assistance in collecting and analyzing the data, and Janice May of SmithKline Beckman for her helpful editorial suggestions.
December 30, 1983
The American Journal of Media:lne
85
ALFRED E. GOLDMAN, Ph.D. SUSAN SCHWARTZ MCDONALD, Ph.D. Philadelphia, Pennsylvania
From National Analysts, Division of &x-Allen & Hamilton, Incorporated, Philadelphia, Pennsylvania. Requests for reprints should be addressed to Dr. Alfred E. Goldman, National Analysts. 400 Market Street, Philadelphia, Pennsylvania 19106.
National Analysts conducted primary research with rheumatoiogists-specificaiiy, two panel discussions, 25 in-depth telephone interviews, and a mail survey of conference (Auranofin Symposium and Workshop) participants-to examine current treatment practices and to probe the rationale and motivations underlying treatment strategies in rheumatoid arthritis. The research identified important areas of consensus in drug perceptions, therapeutic approaches, and disagreements. Physicians differ regarding the mlnfmum time they wait after diagnosing rheumatoid arthritis before initiating remittive therapy, some beginning immediately and others waiting six months or longer. Younger physicians are quicker to initiate remittive treatment than their older colleagues, but both younger and older practitioners are initiating remittive therapy earlier than in the past. Some noteworthy differences between hoepitai-based and office-based practitioners were discerned with resp&t to factors that figure in their decisions to initiate remittive therapy. Differences were also found among physicians in the way they pose drug options to their patients; “authoritarian,” “iibertarian,” and “guided democracy” were names given to the three styles ktentMd. in general, however, physicians report that patients are more directly involved in treatment selection than previously, a trend that may in part be due to the use of more aggressive treatment strategies than in the past and a desire to share the psychoiogk burden of those decisions. Findings suggest that goid w will continue to be a mainstay first-line disease-modifying agent in the treatment of rheumatoid arthritis but that there may be less reluctance to use other agents as physicians become increasingly familiar and comfortable with aiternative options, especially penkiiiamine and immunosuppressive agents. At the request of the Auranofin Symposium executive committee, National Analysts, a marketing research and consulting organization, conducted research on the way clinicians confront rheumatoid arthritis, with the express purpose of sharing the results with the conference participants. The data collected provide a descriptive picture of how practitioners are dealing with rheumatoid arthritis, how they view the available treatment options, and the basis for their decisions in treating individual patients. The mix of techniques selected-“quantitative” procedures (which lend themselves to statistical analyses) and “qualitative” procedures (which address attitude structure rather than frequency)-together provide a measure of underlying motivation as well as a count of clinical practices.
December 30, 1983
The American Journal of Medlclne
79
ORAL GOLD SYMPOSIUM-GOLDMAN
TABLE
I
and MCDONALD
What is Minimum Length of Time You Would Prefer to Wait After a Confirmed Diagnosis of Rheumatoid Arthritis Has Been Made, Before Initiating Ftemittive Therapy? Percent
Reply
No minimum-stat-t
immediately
if warranted
54 24 16 2 2
3 months 6 months 12 months More than 12 months
METHODS
The findings reported here reflect three complementary methods: (1) two group discussions with rheumatologists; (2) approximately 25 in-depth telephone interviews with rheumatologists across the country: and (3) a mail survey of 121 conference participants, 94 percent of whom are rheumatologists. Group Discussions. Group interviews involved eight to 10 practitioners assembled in the presence of a trained moderator. The sessions lasted about two hours, capitalizing on the principles of group dynamics and peer interaction to elicit responses of a depth that would not be forthcoming in a survey. Valued for their sensitivity and candor, group sessions do not provide statistically projectable data; they fall instead under the heading “qualitative research.” Two sessions were conducted, one in Philadelphia and the other in Atlanta. Like group sessions, indiIndividual In-Depth Interviews. vidual telephone interviews are qualitaltive in nature and are not appropriate for statistical projection. They do, however, permit a more exhaustive and intensive conversation and broader geographic representation.
% Who Would Start Remittive Therapy Immediately, If Warranted 75% (61%)
50%
.b)
25%
0% Under 40
40 - 49 PHYSICIAN
50 + AGE
Figure 1. Younger physicians, as a group, tend to initiate remitting drug therapy sooner than their older colleagues.
90
December 30, 1993
The American Journal of Medicine
Mail Survey. All those who accepted the invitationto attend the symposium were sent questionnaires and asked to participate in the survey. Of the nearly 130 who did attend the sessions, 121 (or better than 90 percent) returned questionnaires in time to be processed prior to the symposium. Thus, responses reported here represent a view of clinical practices before exposure to conference material and constitute a virtual “census” of the audience that was present. Because only a few respondents in the survey sample did not describe themselves as rheumatologists, we were unable to compare the responses of different specialty groups. We did, however, make use of another practice variable-office-based versus hospital-based-a classification scheme reflecting the way physicians defined themselves when given those options in the questionnaire. For purposes of analysis, the responses of solo and group (office-based) practitioners were combined. The term “remittive therapy,” although not specifically defined, was equated in the questionnaire with the following: injectable gold, hydroxychioroquine, peniciilamine, and im-
munosuppressive agents. RESULTS Along with some areas of broad consensus, the data suggest noteworthy differences in treatment styles and strategies-variations that seem to reflect the different ways physicians have of resolving the competing values and perils endemic to treatment. Both the group and individual in-depth interviews suggested that rheumatologists share a sense of frustration and vulnerability in treating rheumatoid arthritis, a disease described by one practitioner as “monstrous.” The best they could hope for is to halt or delay progression of the disease with drugs that may or may not work and that pose alarming risks. There was disagreement about when to initiate therapy with a remittive agent, which surfaced in the survey as well as in the qualitative phase. Some begin therapy promptly, as soon as they suspect that a patient has rheumatoid arthritis, because they believe that to wait for evidence of bone erosion is, as one put it, “to lose the game.” Others prefer to wait, sometimes for months, before initiating remittive therapy. As Table I illustrates, 54 percent of the physicians surveyed would start therapy “immediately if warranted” once a confirmed diagnosis had been made, 24 percent would wait a minimum of three months, and another 18 percent would wait six months. Only 4 percent would wait a year or longer. Personal discussions identified at least one important reason for the variation, the possibility of spontaneous remission. Some prefer to temporarily defer treatment, hoping the disease will halt independently. Factors like physician age, recency of training, and institution also seem to affect the speed with which they begin remittive treatment. Younger physicians were quicker to initiate therapy than their older colleagues: 61 percent
and MCDONALD
ORAL GOLD SYMPOSIUM-GOLDMAN
of those under age 40 would initiate therapy immediately, whereas only 53 percent of those in their 40s and 48 percent of those over 40 would make the same decision (Figure 1). Roughly half the respondents surveyed, younger and older practitioners alike, claimed to be initiating remittive therapy earlier than they used to, suggesting that the perils of the disease are outweighing the perils of the treatment in these physicians’ clinical equation. No one is waiting longer to initiate remittive therapy than in the past. The primary factors precipitating a decision to begin remittive therapy are of interest. Physicians were virtually unanimous in emphasizing bone erosion, failure to respond to nonsteroidal anti-inflammatory drugs, and joint swelling as triggers in that decision, but there were some striking differences in the priorities of officebased and hospital-based practitioners (Table II). Office-based physicians cited a wider number of factors, including nonphysiologic ones, as “very important” in this decision to begin remittive therapy. Specifically, more office-based practitioners considered threatened job loss, frequency of flare-ups, curtailment of activities,
50%
1
TABLE II
Mean Importance of Top 10 Factors in Decision to Initiate Remittive Therapy Mean’
Factor Increasing number and severity of swollen joints Not responsive to nonsteroidal anti-inflammatory drugs Presence of bone erosion More frequent flare-ups Presence of other extra-articular manifestations Curtailment of activities Threatened job loss Unrelenting pain Increased fatigue, malaise, weakness Joint space narrowing All others
5.3 5.2 5.2 4.6 4.6 4.5 4.5 4.4 4.2 4.2 <4.0
’ On scale from 1 to 6, where 1 = not at all important and 6 = very important.
and unrelenting pain to be “very important” in making their decision. Figure 2 lists those factors and the percentage who deemed them “very important,” using a 6-point scale. We can only speculate on the reasons for this difference: since hospital-based physicians are, on
THREATENED JOB LOSS
CURTAILMENT OF ACTIVITIES 50%
40
.., . .,..... ::::: L I
30
(25%)
1
20
10
0
Office Based
Hospital Eased
::::::::::::::::::: ::::::::::::::::::: ......,...,....,. .::::::.:::::::::::: .....).
:::::>..:.:.:.:.:.: ::::::y::::::::: (10%) :::y$:::::::y: :::>.::::y::$;>, :::y:::::::::::> ::::;::::::::::::: ::::::::::::::.:.: ......#.......L..
Office Based
UNRELENTING PAIN 50%
Hospital Based
MORE FREQUENT FLARE UPS
50%
1
:>.~.~.~.~.>.~.~. ::::::::::::::::::: >:s’c.:.:.:.:.>: :y::::::::::::::: .‘.....S........‘.
1
40.
30 -
30
(27%)
-
20
Figure 2. Importance of factors in decision to initiate remittive therapy for office- and hospital-based physicians. (Percent citing factor as ‘very important, ’ or ‘6’ on a 6 point scale.)
10
0 b
-
Office Eased
Hospital Eased
December 30, 1983
(22%) .. ::::::::::::j:::: ::::::::;::::::::: ..::::_.:.~.. ..........~...~.~.~.~.. ......... :::::::::::::::::: .. . (10%) :.:.:.:.:.:.:.:.:.
. . . :.:.:.;.:.:.:.:.:. :::::::::::::::::: :::::::::::::::::: .:.:.:.:.:.>:.:.: ;:>i:):$:>;:;: Office Based
... ..,.. :::::::::::y:::: .:+>:.:.:.:.:.: ::y::::$.:.:. :...>:.:.::::..:..: .>:.>:.:.:.:.:.:
Hospital Based
The American Journal of Medicine
81
ORAL GOLD SYMPOSIUM-GOLDMAN and MCDONALD
TABLE Ill
Characteristics Associated with Key Remittlve Agents (Percent Assigning Description to Each Drug) Percent
Characteristic
Injectable Gold Most Most Most Most
effective generally effective for aggressive disease effective for smouldering disease preferable for young adults
90 a4 79 78
Hydroxychloroquine Least costly Causes least frequent side effects Least effective generally Causes fewest serious side effects
96 96 90
88
Penicillamine Most difficult to tolerate over long period Causes most frequent side effects Causes most serious side effects
average, slightly less aggressive
80
78 76
in initiating remittive
therapy, they may, in fact, be less responsive to what they believe are more “peripheral” considerations. It is also possible that office-based physicians know their patients better and that the intimacy or continuity of that relationship makes patients’ physical discomfort and psychologic stress seem more urgent. There was general agreement about key characteristics of first-line remittive agents, particularly such
features as efficacy, toxicity, and safety. The vast majority designated injectable gold as “most effective generally” for either aggressive or smouldering disease, most preferable for young adults, and most costly (Table Ill). At least eight in 10 termed hydroxychloroquine most convenient and safest (causing least frequent and fewest serious side effects) but also least generally effective. Finally, most viewed penicillamine as least safe and most difficult to tolerate over a long period of time. Consistent with those drug profiles, most of the survey respondents initiate remittive therapy with gold in the majority of patients. Table IV reports the average percentage of patients in whom gold, hydroxychloro-
TABLE IV
Initiating Drugs for Remittive Therapy
Percent of MDs ever starting patients on Injectable gold Hydroxychloroquine Penicillamine Immunosuppressive
agents
100% 88 % 76% 27%
Average percentage of patients receiving remittive therapy who were started on Injectablegold Hydroxychloroquine Penicillamine Immunosuppressive agents
71% 16% 14% 2%
82
December 30, 1883
The American Journal of Medicine
quine, and penicillamine therapy was started. However, some important variations in treatment style and drug perspectives emerged on further examination of the figures. Because physicians perceive treatment with remittive agents to be fraught with idiosyncracy and risk, and because patients often have strong feelings about the mode of therapy, a rigid progression of drugs or fixed treatment hierarchy could not be identified. Physicians who participated in the qualitative phase described a number of relevant considerations. The disease stage and rate of progression, the patient’s age and cardiovascular condition, and his or her mental set (especially reliability and anticipated compliance) were all factors that militated against one agent and for another. For example, some patients were more fearful of gold injections than of the serious, even life-threatening, side effects that oral alternatives might precipitate. That fear could be decisive in eliminating gold therapy as a serious option. By contrast, some patients came to their doctors expecting gold shots and that expectation guided the clinician’s course of action. Some physician variables also were identified. In particular, three distinct physician orientations to treatment were encountered in the qualitative analysis. One approach, labeled “authoritarian,” offers a decisive course of treatment, saying to the patient, “This is the drug I want you to take, and these are my reasons.” The second, more “libertarian” approach offers the patient a menu of acceptable options; it points out the strengths and the liabilities of each one and then says, “Now you may choose.” The third is perhaps best described as “guided democratic.” This approach is similar to the libertarian; it offers the patient a series of options, outlines the advantages and risks attached to each, but then communicates more or less directly the physician’s preference. Like many classification schemes, this one should be viewed as a continuum rather than a set of rigid categories. The authoritarian model is apt to yield under patient pressure, and the libertarian model may be a rigged game in which the physician successfully imposes his preference, however subtly. We believe that these differences can have important implications for therapy. Indeed, the survey of symposium participants provided some data to substantiate or quantify the typology. The majority (64 percent) reported that they “tend to select a single drug and present that to the patient,” whereas the rest indicated that they “tend to present the patient with two or more options.” The less structured (qualitative) discussions with physicians suggest that the tendency to involve patients more directly and explicitly in the decision-making process is growing.
ORAL GOLD SYMPOSIUM-GOLDMAN
This trend may be in part a reflection of more aggressive use of remittive agents and first hand familiarity with a wider array of options. It is consistent with movement toward more fastidious informed consent in treatment protocols. More importantly, it is a mechanism for sharing some of the burden in the decisionmaking process. Moreover, as many physicians pointed out in the in-depth interviews, rheumatoid arthritis therapy is really a lifetime collaborative effort between patient and physician; without the wholehearted commitment and compliance of the patient, success of the treatment is seriously compromised. Involvement of the patient in choosing among therapeutic options may be an attempt to enlist the patient’s commitment to the elected course of treatment. Practitioners differ from one another in other important ways. Although all are privy to the epidemiologic data about incidence and severity of drug side effects, they come to varying conclusions about the risks actually posed to their own patients and about the degree to which those risks are tolerable. The differences reflect variations in the way physicians weigh statistics versus clinical perceptions-in other words, how they assess “actual” risk in relation to “perceived” vulnerability. A classic illustration of the difference between statistical and perceived risk is the fear occasioned in some people by airplane travel, even though they recognize that, statistically speaking, the car trip to the airport is more dangerous. In a similar vein, nearly all the physicians surveyed agree that hydroxychloroquine is generally the safest drug, yet some prefer to use penicillamine because the small risk of permanent blindness from hydroxychloroquine is unacceptable either to them or to their patients, especially given what some physicians view as the questionable efficacy of that drug. Others would prefer to play these very complex odds another way; with careful monitoring, they are entirely comfortable in the belief that major problems with hydroxychloroquine can be detected and reversed. Still others (27 percent of the sample) initiated remittive treatment of a select handful of patients (usually fewer than 5 percent) with immunosuppressive agents, whereas the rest of our survey respondents rejected that level of risk (Table IV). Clinical experience, residency training, institutional CUStOfW personality, patient demographics, and a host of related factors influence those assessments; as physicians made clear in every phase of our research, every new patient invites some recalculation of the odds. In fact, besides providing information on physicians’ treatment strategies, the survey also provided an opportunity to compare approaches in the management of several “case studies.”
MANAGEMENT
and MCDONALD
PROBLEMS
Case 1. Physicians were asked how they would respond to evidence of an increase in bone erosion and cartilage loss in a 45year-old woman with moderately aggressive rheumatoid arthritis who was experiencing intermittent flare-ups affecting mostly the hands, wrists, and ankles. All but one of the respondents were prepared to treat this patient with remittive agents and, of those, 94 percent recommended injectable gold. Nearly half-approximately 46 percent-would offer the patient several options. Almost four in 10 would offer penicillamine, and more than a quarter would also recommend hydroxychloroquine. Two were prepared to initiate immunosuppressive therapy. (Note that in Figures 3 through 5, which illustrate the case study responses graphically, the encircled drugs are those that would be offered to the patient along with the percentage who would recommend each drug. More than one drug could be selected if physicians would choose to present several options. The modal daily starting dose is linked to each drug by an arrow.) There was a fair degree of consensus in the starting dosages for any given drug: nine in 10 would give 50 mg or more per week of gold, six in 10 would recommend 400 mg of hydroxychloroquinine (except for injectable gold, all drug dosages cited are per diem.), and seven in 10 would begin with 250 mg of penicillamine. The remainder would initiate therapy at more conservative dosage levels. This represents a fair degree of consensus because, from a statistical point of view, large majorities indicated that they would use the same dosage. On the other hand, there was a sizeable minority in disagreement. Indeed, our group discussions
88% 50mg.
Figure 3. Treatment strategies in a 45-year-old woman (Case 1) with moderately aggressive rheumatoid arthritis, intermittent flare-ups, increase in bone erosion and cartilage loss.
December 30, 1983
The American Journal of Mediclne
83
ORAL
GOLD
SYMPOSIUM-GOLDMAN
and MCDONALD
n
PBNICIUJMINE
MAXIMUM
F@re 4.
Treatment strategies on an engineer (Case 2) with
mild nagging joint symptoms and significant ear/y morning stiffness; currently receiving prednisone (5 mg per day) and
aspirin.
J
Figure 5, Treatment strategies for a 33-year-old woman (Case 3) with aggressive rheumatoid arthritis: no improvement with gold, visual symptoms with hydroxychloroquine.
INJECTABLE GOLD
HYDROXYCHLOROQUJNE
% Rspa-ting
Increased UseOf:
PENICILLAMINE
IMMUNOSWPRESSIVES
75% -
75% 7 Ki%I
50%.
PENICILLAMINE
IMMUNOSUPPRESSIVES
Undw40
40.49
PHYSICIAN AGE
1
50+
Under 40
40.49
50+
PHYSICIAN AGE
Figure 6.
Percentage of physicians reporting increased use of remittive agents.
F&e 7. Relationship between physician age and changing drug usage. Older physicians are using more penicillamine and fewer immunosuppressive agents than younger ones.
suggest that practitioners treating rheumatoid arthritis often assume greater consensus than really exists. Case 2. There was considerably less agreement as to whether a remittive agent was appropriate for the second patient, an electrical engineer with mild, nagging joint symptoms and significant early morning stiffness. One fifth of the physicians indicated that this patient was not yet a candidate for remittive therapy. Those who did prescribe a remittive agent were most likely to recommend gold (84 percent) or hydroxychioroquine (47 percent). Only about 20 percent thought that peniciliamine was warranted. There was consensus on the appropriate starting dose for gold (88 percent would
recommend 50 mg or more) but less agreement on dosage strategy for hydroxychioroquine. Case 3. The third case drew near unanimity: 91 percent were prepared to recommend peniciiiamine to this 33-year-old woman with aggressive rheumatoid arthritis, who had not experienced any improvement with gold therapy and had had visual symptoms with hydroxychioroquine. Another 44 percent were prepared to recommend immunosuppressive agents, with preference divided roughly evenly between azathioprine and methotrexate. The modal daily starting dose for azathioprine was 50 mg, and approximately 1 mg for methotrexate. (Note: Many physicians appear to be
84
December 30.1383
The American Journal of Medlclne
ORAL GOLD SYMPOSIUM-GOLDMAN
and MCDONALD
splitting methotrexate dosages and/or prescribing an alternate-day or a weekly regimen.) Many, however, were willing to exceed 100 mg of azathioprine (37 percent) or 3 mg of methotrexate (40 percent) per day (Figure 5). These numbers offer a synopsis of current practices, and a possible forecast of future treatment strategies. There is some evidence to suggest that physicians are increasingly receptive to the use of immunosuppressive agents. They are also using more penicillamine than in the past. Figure 6 indicates percentages of the survey sample reporting either less frequent or more frequent use of all major remittive agents; gold and hydroxychloroquine are, more or less, holding their own. Of particular interest is the fact that younger physicians are most apt to increase their use of immunosuppressive agents, whereas older physicians are reporting growing use of penicillamine (Figure 7). Those figures forecast the near-term, and perhaps long-term, trend in the treatment of rheumatoid arthritis: physicians over 40, having had substantial experience in their practices with penicillamine, are becoming less leery of that drug; their colleagues under 40, fresh from closer contact with immunosuppressive agents in their training, are discounting or accepting some of the reputed dangers of those agents.
similar data against which to compare these findings exist, and there is no way of judging what the responses to these questions might have been several years earlier. These data represent a “snapshot” taken at one point in time depicting a selected universe. It is tempting, however, to extrapolate and suggest that remittive agents like penicillamine and immunosuppressives will be used increasingly as physicians take a more aggressive posture in treating rheumatoid arthritis. Admittedly, the younger physicians who indicate greater willingness to prescribe immunosuppressive agents may, with age, grow to act more like their older, more cautious colleagues. On the other hand, familiarity may breed some comfort, if not contentment, with the second- and third-tier drugs, provided nothing occurs to dramatically alter the risk scenario or obviate the need for these agents. If there is one conclusion that can be safely drawn, it is a familiar one: reasonable practitioners will continue to differ in the treatment strategies they devise for rheumatoid arthritis. Those differences are presently bound by shared constraints and anxieties. Future treatment for rheumatoid arthritis is likely to remain a careful, vigilant course between the perils of the disease and the continuing perils of treatment.
COMMENTS
ACKNOWLEDGMENT
One of the misfortunes of medical survey research is that it is seldom published where physicians will encounter it because it is not designed to educate physicians about their own behavior. To our knowledge, no
We would like to thank Maria Figueroa of National Analysts for her assistance in collecting and analyzing the data, and Janice May of SmithKline Beckman for her helpful editorial suggestions.
December 30, 1983
The American Journal of Media:lne
85