The Importance of Physician Presence in Nursing Homes for Residents with Dementia and Pneumonia

Article Outline

• Abstract
• Methods
  • Measures
  • Analysis
• Results
• Discussion
• References
• Copyright

Objective

To study whether physician presence in the nursing home is related to clinical decision making, certainty, and honoring care preferences for patients with dementia and pneumonia.

Design

Cross-sectional survey of physicians.

Setting

Nursing homes in the United States and the Netherlands.

Participants

Twenty-four US and 38 Dutch physicians who provide care for nursing home patients.

Measurements

Physicians reported their presence in the nursing home, diagnostic and treatment decisions for patients with dementia who had pneumonia, certainty about the diagnosis and patient and family preferences, and the extent to which they honored these preferences. These variables were examined in reference to physician presence in the nursing home.

Results

Physicians with higher nursing home presence were less likely to order a chest x-ray and to hospitalize patients with dementia and pneumonia, although this difference was not significant when adjusted for country. They also were more likely to be certain of family preferences, a difference that held even when adjusted for the strong confounder of country.

Conclusion

Physician presence in the nursing home relates to some treatment decisions for patients with dementia and pneumonia. Policies that affect physician presence may change health care practices and related costs. Future studies should more closely examine how physicians use their time so as to better understand the importance of presence and what the US health care system might learn from the Dutch system.

Keywords: Nursing homes, physicians, decision-making, hospitalization, pneumonia

Nursing homes (NHs) provide care to older adults with increasingly complex health care needs—a role previously filled by hospitals. Because of the rising acuity of NH patients and the growing older population, NHs will play an increasingly important role in health care. It may be surprising, then, that NH physicians have been described as “missing in action,” a common concern of families who have a relative in a NH and are unclear as to what medical care is indicated.¹, ² Physicians also may be uncertain as to what care is indicated and what the family prefers, especially when they spend little time in the nursing home. In these cases, their judgment may be highly discretionary with the tendency to act under the assumption that “more is better.”³, ⁴ This uncertainty may affect clinical decisions including transfer to the hospital.

To test if physician presence in the NH is related to decision making, it is necessary to study a sample of physicians with variability in physician presence and then to measure decisions regarding a common clinical scenario. Physicians in the United States constitute a cohort with limited presence, as NH practice is only 4% of work time among the 20% of physicians who practice in an NH.⁵ In the Netherlands a smaller percentage (4%) of the country's physicians work in NHs (Mark Siekerman, Huisarts, Verpleeghuisarts en arts voor verstandelijk gehandicapten Registratie Commissie [Dutch physician registration commission], personal oral communication, 2010), but these physicians have completed a 3-year residency training program in NH medicine, work almost exclusively at the NH as their site of practice, and are part of a team of physicians at the NH, at least one of whom is present on a daily basis.⁶ Thus, this study includes physicians from both the United States and the Netherlands. The clinical scenario is one of an NH patient with dementia who became acutely ill with pneumonia. Although Osler more than a century ago called pneumonia “the friend of the aged,” referring to a quick and not very painful death,⁷ recent studies report that such patients can suffer significantly.⁸ The benefits of aggressive care to lengthen life in patients suffering from terminal dementia are also questioned.⁹ The lack of consensus among both physicians and family members on the medical effectiveness and the ethical appropriateness of treating to prolong life in patients with dementia leads to high variability when it comes to treatment decisions, which is why this common clinical scenario was chosen for study.

The aim of this study was to explore the hypotheses that physicians who are highly present in the NH (1) make different diagnostic and treatment decisions, (2) are more certain of their decisions and of patient and family preferences, and (3) are more likely to honor those preferences, than physicians who are less present.

Back to Article Outline

Methods 

The study population was composed of a cross-sectional sample of physicians in the United States and the Netherlands who treat NH patients. In the United States, 2 strategies were used to recruit physician respondents. First, 12 NHs (4 from each of 3 states) were randomly selected in Iowa, New Jersey, and North Carolina, 3 states that exhibit a range of characteristics related to patterns of physician visits. Specifically, New Jersey exhibits a high number of physician NH facility visits (only 28% of facilities reported no physician visits in the past 14 days), North Carolina a mid-range number (52% of facilities reported no physician visits in the past 14 days), and Iowa a low number (65% of facilities reported no physician visits in the past 14 days).¹⁰ In these facilities, an NH staff member was contacted monthly and asked to identify physicians who had recently treated a case of pneumonia for a patient with dementia and to ask those physicians to complete the questionnaire on that patient. Physicians affiliated with geriatric training programs in North Carolina were also surveyed. In total, 106 US physicians received a mailed survey beginning in October 2006. A reminder postcard was sent to nonrespondents 2 weeks after the initial mailing; continued nonrespondents were contacted by telephone.

In the Netherlands, the survey was mailed to 106 physicians beginning in July 2006. These physicians worked in 27 NHs that had been randomized into the control group of a study on using a prognostic score in the management of pneumonia. The physicians were asked to complete the survey for their next patient with dementia who was diagnosed with pneumonia. Two reminder phone calls were made to the contact person at each facility if the physicians did not return the completed survey in a timely manner.

Study methods were approved by the Institutional Review Board of the University of North Carolina at Chapel Hill and the Medical Ethics Committee of VU University Medical Center, Amsterdam.

Measures 

Drawing from the literature on uncertainty in physician decision making,¹¹ questions were developed that measured diagnostic and treatment characteristics of the pneumonia; certainty of diagnosis, prognosis, and patient and family preferences; and behaviors related to honoring patient and family preferences. Data also were collected about physician characteristics such as age, years of experience, whether or not he or she provides hospital care, and presence in the nursing home as percentage of practice time spent in a NH and frequency of visits to individual patients. Respondents were asked questions about the patient including age, illness severity (measured on a 9-point Likert scale from 1 [not sick] to 9 [moribund]),¹² number of symptoms (reported from a list of 12 symptoms common in pneumonia), and dementia severity using the Bedford Alzheimer Nursing Severity Scale.¹³, ¹⁴

Analysis 

To assess physician presence in the NH, a summary NH presence score was developed. The scale assigned points based on the percentage of practice time spent in the NH and the frequency of visits to a typical NH patient. The score ranged from 0 to 12, with lower scores denoting lower physician presence.

Survey data were double-entered into a Microsoft Access database, imported into SAS 9.1 (SAS Institute, Inc., Cary, NC), and checked for accuracy and logical errors, with discrepancies resolved by reference to the original form. Analyses compared distributions of variables by country, and also the relationship between physician presence and diagnostic and treatment decisions, certainty, and honoring patient and family preferences. Because of the small sample size, we were interested in trends (P < .10). Frequencies and nonparametric statistical tests (to account for small sample size) were computed. For continuous data, the Wilcoxon nonparametric test was used; for categorical data, the Fisher's Exact test and logistic regression (using the exact option) were used.

Back to Article Outline

Results 

In the United States, 24 completed surveys were received, which, after excluding 18 known ineligible cases (eg, the physician had not treated an eligible patient case) and 10 presumed ineligible cases (ie, the percentage of known ineligibles extrapolated to the number of nonrespondents) from the original 106 mailed surveys, represented a 31% response rate. The average time between the patient's illness and the US physician completing the questionnaire was 47 days (range 1 to 171 days). In the Netherlands, 38 surveys were received, representing a 36% response rate. Thirty-seven percent of the Dutch physicians completed the survey within 1 day, with the mean being 12 days (range 1 to 95 days). The main reason for not completing the questionnaire was not having a case of pneumonia during the period of study.

Physicians in both countries were similar regarding age and years of practice, although there were somewhat more females in the Dutch sample (68% versus 42%; P = .06) (Table 1). The US physicians were about equally divided between being family physicians and internists, whereas all of the Dutch physicians listed NH medicine as their specialty. Most US physicians (75%) reported they provide care to their NH patients if they are hospitalized, whereas none of the Dutch physicians did so. Many US physicians (46%) also reported they make more money if the patient is hospitalized. Dutch physicians spent significantly more time in the NH, with 82% spending more than 75% of their practice time at the NH in contrast to 88% of US physicians who spent less than 50% of their practice time there (P < .001). Dutch and US physicians did not differ significantly in the frequency with which they visit individual NH patients, although there was a trend (P = .11) for Dutch physicians to visit more often. Dutch physicians had significantly higher NH presence scores than US physicians (9.5 versus 6.0; P < .001), the difference mostly accounted for by percentage of time spent at the NH.

Table 1. Demographic and Practice Characteristics of Respondents

	Frequency	Percent	Frequency	Percent
Gender
Female	10	42	26	68	.06‡
Medical specialty
Family Medicine	13	54	—	—	—
Internal Medicine	11	46	—	—
Nursing Home Medicine	—	—	38	100
Provide hospital care for nursing home patients (yes)	18	75	0	0	<.001‡
Make more money if patient hospitalized (agree or strongly agree)	10	46	0	0	<.001‡
Practice time in nursing home
<50%	21	88	2	5	<.001‡
50%–74%	1	4	5	13
75%–90%	1	4	14	37
>90%	1	4	17	45
Visit individual nursing home patients
Daily	0	0	6	16	.11‡
Weekly	5	21	6	16
Monthly	11	46	19	51
Every 2 months	8	33	6	16
< Every 2 months	0	0	0	0

Patient cases in both countries were similar regarding dementia severity and severity of illness (Table 2). Dutch patients, however, were reported by their physicians to have significantly more clinical symptoms of pneumonia than reported by US physicians about their patients (6.8 versus 4.2; P < .001). Also, patients in the United States had resided in the NH longer (3.9 years) than those in the Netherlands (1.8 years; P = .01) and had a longer relationship with their physician (4.1 years) than the Dutch patients (1.2 years; P < .01).

Table 2. Demographic, Diagnostic, and Treatment Characteristics of Patient Cases

	Frequency	Percent	Frequency	Percent
Gender
Female	15	63	22	58	.79†

With respect to NH presence and diagnostic and treatment decisions, physicians who ordered a chest x-ray had significantly lower NH presence scores than those who did not (6.4 versus 9.0; P < .001), yet there was no difference in the certainty of the pneumonia diagnosis based on NH presence (8.1 versus 8.0; P = .90; see Table 3). Also, presence was significantly less for physicians who hospitalized patients (5.5 versus 8.9; P < .001), and treated with oxygen (7.1 versus 9.2; P < .001) and rehydration (6.0 versus 8.7; P < .001). Further, there was a trend for physicians who used bronchodilators and ordered tube feeding to have lower NH presence scores. The decision to treat with antibiotics was not associated with physician presence.

Table 3. Relationship between Physician Nursing Home Presence and Decision-Making, Certainty, and Honoring Preferences

	Presence Score, Mean (SD)∗		OR (95% CI) †
	‘No’ or ‘Disagree or Strongly Disagree’ Response	‘Yes’ or ‘Agree or Strongly Agree’ Response	Unadjusted	Adjusted for Country
Diagnostic and Treatment Decisions
Ordered a chest x-ray	9.0 (1.9)	6.4 (1.9)	1.9 (1.4–2.7)	1.1 (0.6–1.8)
Hospitalized patient	8.9 (1.9)	5.5 (1.2)	2.7 (1.6–5.5)	1.8 (0.9–3.9)
Treated with oxygen	9.2 (2.0)	7.1 (2.0)	1.6 (1.2–2.3)	1.2 (0.8–1.8)
Treated with rehydration	8.7 (2.1)	6.0 (1.6)	1.9 (1.3–2.8)	1.2 (0.7–2.1)
Treated with bronchodilators	8.6 (2.1)	7.3 (2.4)	1.3 (1.0–1.7)	1.1 (0.8–1.7)
Treated by tube feeding	8.3 (2.2)	6.2 (2.2)	1.5 (1.0–2.5)	0.9 (0.4–1.8)
Treated with antibiotics	9.0 (1.4)	8.0 (2.3)	1.2 (0.9–1.8)	0.9 (0.5–1.5)
Certainty
Certain of diagnosis	8.0 (2.2)	8.1 (2.3)	1.0 (0.7–1.3)	1.3 (0.8–2.2)
Certain of patient preferences	7.6 (2.3)	8.4 (2.2)	0.9 (0.7–1.1)	0.8 (0.6–1.2)
Certain of family preferences	6.8 (3.0)	8.4 (2.1)	0.7 (0.5–1.0)	0.5 (0.2–0.9)
Honoring Patient and Family Preferences
I honored patient preferences	8.5 (2.0)	8.0 (2.3)	1.1 (0.8–1.5)	1.0 (0.6–1.5)
I honored family preferences	8.0 (2.4)	8.2 (2.2)	1.0 (0.7–1.3)	0.6 (0.3–1.0)
I made a decision not in keeping with family preferences	7.9 (2.2)	10.2 (1.5)	0.5 (0.2–1.0)	0.6 (0.3–1.4)

Odds ratios with a confidence interval that excludes the value of 1 are shown in bold. These bold items have P < .05, whereas items with P < .10 are bold and italicized.

OR, odds ratio; CI, confidence interval.

Before adjustment for country, physician presence was not significantly related to certainty or tendency to honor patient and family preferences. Still, physicians who reported that they made a treatment decision that was not in line with the family's preferences had higher NH presence than those who did not deviate from family preferences (10.2 versus 7.9; odds ratio 0.5, P < .10).

Because physician NH presence was highly correlated with country, Table 3 also shows all results adjusted for country. The only difference that was significantly related with presence was that physicians with higher presence were more certain of family preferences (8.4 versus 6.8; P < .05). There also was a trend for these more present physicians to honor family wishes (P < .10).

Both gender and physician age were examined as possible confounders, and in all cases these variables did not affect the significance of the findings.

Back to Article Outline

Discussion 

This study found that physicians who are more present in the NH make some different decisions for their patients with dementia and pneumonia than those who are less present, although other country differences also may account for these findings. Also, those who are more present are more certain of family preferences, even when controlling for the strong country effect exerted in this sample of physicians from the United States and the Netherlands.

Dutch and US physicians did not differ significantly in the frequency with which they visited individual NH patients. This means that the differences in the NH presence score were attributable to the marked variation in the amount of practice time that the physicians spend at the NH. The NH presence score therefore likely captures the broader notion that when physicians are more frequently physically present, it may be easier to visit a patient immediately when a crisis develops, and to be generally more available to the nursing home staff. More practice time at the NH also allows for more opportunity to interact with families, and better knowledge of the staff and their capabilities.

Physicians who are less present in the NH were significantly more likely to hospitalize their patients (odds ratio 2.7; P < .001), but this relationship was not significant when adjusted for country. Although Dutch physicians are much more present in the NH and other literature confirms that they hospitalize less than 0.6% of NH patients with dementia during an acute episode of pneumonia,¹⁵ this difference could be because of features of the 2 countries' health care systems rather than physician presence in the NH per se. That is, US physicians work in a fee-for-service system, whereas Dutch physicians are primarily salaried; further, 46% of the US physicians reported that they make more money if the patient is hospitalized.

These findings suggest that disparities in hospitalization rates between the 2 countries are related to societal differences in physician training and the organization of NH care,¹⁶ as well as to physician presence. Physicians who are not as present likely have duties elsewhere, including in the hospital. It is thus not surprising to find them more comfortable with a hospital-oriented approach to care. Physician presence did not relate to the use of antibiotics, which is an important finding considering that studies show that NH residents with pneumonia do just as well when treated with oral antibiotics in the NH.¹⁷, ¹⁸

On the other hand, physicians who were less present in the NH were significantly more likely to order a chest x-ray, and to treat with oxygen, and rehydration; they also were marginally more likely to treat with tube feeding and bronchodilators. None of these differences were significant when adjusted for country, however, meaning that these approaches are attributed to the higher use of these technologies in the United States.

Dutch physicians reported significantly more patient symptoms than the US physicians. This may be because they are more present and more likely to see symptoms, or that the nurses more readily report symptoms to physicians because they see them more. It may also be that because the Dutch physicians tended to have completed the questionnaire sooner after the case than the US physicians (mean 12 versus 47 days, respectively), they recalled more details about symptoms.

When adjusted for country, physicians who had less NH presence were significantly less likely to feel certain of family preferences (odds ratio 0.5, P < .05), which suggests that presence allows physicians to develop a relationship with the family that includes a discussion of care goals. Being less aware of preferences may encourage physicians toward more aggressive treatment when preferences to the contrary are not known. Further, less familiarity may be why these same physicians were marginally less likely to honor family preferences (odds ratio 0.6, P = .06), a difference that was not related to country but to presence.

Ensuring quality care for people who reside in NHs, including decreasing hospital transfer, is a central issue for care providers and policy makers.¹⁹ Many hospital transfers are potentially avoidable or inappropriate.²⁰ Factors that influence the decision to hospitalize NH residents include availability of diagnostic services, adequacy of trained staff, pressure from staff or family, nurse-physician communication, and the financial and emotional cost of transfer.²¹ Of note, in one study, physicians examined only 30% of hospitalized nursing home patients before the decision was made to transfer them to the hospital.²² An increase in physician presence in the NH could affect many of these factors in a way that might reduce hospitalization. Shared decision making can be enhanced with communication that allows the physician to offer expertise, including the potential lack of benefit of hospitalization. Presumably this would be well accepted, as families want more communication.²³ This might also reduce the wide variation in physician practice style that is seen in the decision to hospitalize, even among physicians who work in the same setting.²⁴

Although there are limits to making comparisons between a small, homogeneous country such as the Netherlands and the enormous, complex, and diverse system in the United States, the Dutch health care system is worth studying toward the end of reduced hospitalization and improved care. US policy makers are already studying the way the Netherlands finances its health care²⁵ and the influential American Association of Retired Persons (AARP) ranks the Netherlands as the country that takes the best care of its older citizens.²⁶ Policy decisions that support more physician presence in the NH may be important in meeting the health care needs of an aging society and will require innovations in reimbursement, health care delivery systems, and residency training. Presently, US physicians in training have limited interest in providing NH care.²⁷

As a small exploratory project, this study has limitations including limited power to demonstrate associations. In addition, the Dutch physicians generally completed the questionnaire sooner after the patient's illness compared with the US physicians, which introduces the possibility of recall bias and could explain why Dutch physicians reported more symptoms. Further, the measurement of presence was based on self-report, and observational or other methods would be beneficial to more decisively determine the amount of physician presence and how their time is spent. We also do not have information on the presence of nurse practitioners in the NH, who allow the physician to be less present while still having a trained geriatric clinician on site; this is a limitation that affects understanding only of the US system of care, however, as the Netherlands relies on physicians in the NHs. As another concern, because only 1 of 13 relationships between physician presence and decision making, certainty, and honoring preferences is significant after adjusting for country, there is the possibility of a type 1 error (ie, stating that there is a difference when there is not). However, adjusted associations of presence with 5 of the 6 items related to honoring and certainty of family and patient wishes were all in the same direction (odds ratios 1 or smaller), tempering this concern. Another limitation is a low response rate, which is not uncommon for physician surveys and limits generalizability.²⁸, ²⁹ Still, the fact that in this small study there were differences that reached or approached significance suggests that the importance of physician presence is worthy of further examination.

Back to Article Outline

References 

1. Shield RR, Wetle T, Teno J, et al. Physicians “missing in action”: Family perspectives on physician and staffing problems in end-of-life care in the nursing home. J Am Geriatr Soc. 2005;53:1651–1657
   • View In Article
   • MEDLINE
   • CrossRef
2. Hanson LC, Danis M, Garrett J. What is wrong with end-of-life care? Opinions of bereaved family members. J Am Geriatr Soc. 1997;45:1339–1344
   • View In Article
   • MEDLINE
3. Wennberg J, Barnes B, Zubkoff M. Professional uncertainty and the problem of supplier-induced demand. Soc Sci Med. 1982;16:811–824
   • View In Article
   • MEDLINE
   • CrossRef
4. Carter MV. Variations in hospitalization rates among nursing home residents: The role of discretionary hospitalizations. Health Serv Res. 2003;38:1177–1206
   • View In Article
   • MEDLINE
   • CrossRef
5. Katz PR, Karuza J, Kolassa J, et al. Medical practice with nursing home residents. Results from the National Physicians Professional Activities Census. J Am Geriatr Soc. 1997;45:911–917
   • View In Article
   • MEDLINE
6. Hoek JR, Ribbe MW, Hertogh CM, van der Vleuten CP. The specialist training program for nursing home physicians: A new professional challenge. J Am Med Dir Assoc. 2001;2:326–330
   • View In Article
   • Full Text
   • Full-Text PDF (72 KB)
   • CrossRef
7. Osler W. Lobar pneumonia. In: Principles and Practice of Medicine Designed for the Use of Practitioners and Students of Medicine. 3rd ed.. Edinburgh: Young J. Pentland; 1898;p. 108–137
   • View In Article
8. van der Steen JT, Ooms ME, van der Wal G, Ribbe M. Pneumonia: The demented patient's best friend? Discomfort after starting or withholding antibiotic treatment. J Am Geriatr Soc. 2002;50:1681–1688
   • View In Article
   • MEDLINE
   • CrossRef
9. Volicer L. Hospice care for dementia patients. J Am Geriatr Soc. 1997;45:1147–1149
   • View In Article
   • MEDLINE
10. Centers for Medicare and Medicaid Services. MDS Frequencies Report: January 15, 2001. Available at: http//www2.cms.hhs.gov/states/mdsreports/freq3.asp?var=P7&date=3. Accessed October 4, 2005.
    • View In Article
11. Gerrity MS, DeVellis RF, Earp JA. Physicians' reactions to uncertainty in patient care: A new measure and new insights. Med Care. 1990;28:724–736
    • View In Article
    • MEDLINE
    • CrossRef
12. Charlson ME, Sax FL, MacKenzie CR, et al. Assessing illness severity: does clinical judgment work?. J Chronic Dis. 1986;39:439–452
    • View In Article
    • MEDLINE
    • CrossRef
13. Volicer L, Hurley AC, Lathi DC, et al. Measurement of severity in advance Alzheimer's disease. J Gerontol. 1994;49:M223–M226
    • View In Article
    • MEDLINE
14. Bellelli B, Frisoni GB, Bianchetti A, Trabucchi M. The Bedford Alzheimer Nursing Severity Scale for the demented: Validation study. Alzheimer Dis Assoc Disord. 1997;11:71–77
    • View In Article
    • MEDLINE
    • CrossRef
15. van der Steen JT, Ooms ME, Ader HJ, et al. Withholding antibiotic treatment in pneumonia patients with dementia. Arch Intern Med. 2002;162:1753–1760
    • View In Article
    • MEDLINE
    • CrossRef
16. Mehr DR, van der Steen JT, Kruse RL, et al. Lower respiratory infections in nursing home residents with dementia: a tale of two countries. Gerontologist. 2003;43:85–93
    • View In Article
    • CrossRef
17. Kruse RL, Mehr DR, Boles KE, et al. Does hospitalization impact survival after lower respiratory infection in nursing home residents?. Med Care. 2004;42:860–870
    • View In Article
    • MEDLINE
    • CrossRef
18. Naughton BJ, Mylotte JM. Treatment guideline for nursing home-acquired pneumonia based on community practice. J Am Geriatr Soc. 2000;48:82–88
    • View In Article
    • MEDLINE
19. Committee on Improving Quality in Long-Term Care . Institute of Medicine. Improving the Quality of Long-Term Care. Washington, DC: National Academy Press; 2001;
    • View In Article
20. Saliba D, Kington RS, Buchanan J, et al. Appropriateness of the decision to transfer nursing facility residents to the hospital. J Am Geriatr Soc. 2000;48:154–168
    • View In Article
    • MEDLINE
21. Kayser-Jones JS, Wiener CL, Barbaccia JC. Factors contributing to the hospitalization of nursing home residents. Gerontologist. 1989;29:502–510
    • View In Article
    • MEDLINE
    • CrossRef
22. Brooks S, Warshaw G, Hasse L, Kues JR. The physician decision-making process in transferring nursing home patients to the hospital. Arch Intern Med. 1994;154:902–908
    • View In Article
    • MEDLINE
23. Port CL, Zimmerman S, Williams CS, et al. Families filling the gap: Comparing family involvement for assisted living and nursing home residents with dementia. Gerontologist. 2005;45:87–95
    • View In Article
    • CrossRef
24. Cohen-Mansfield J, Lipson S. To hospitalize or not to hospitalize? That is the question: An analysis of decision making in the nursing home. Behav Med. 2006;32:64–70
    • View In Article
    • MEDLINE
    • CrossRef
25. Enthoven AC, van de Ven WP. Going Dutch—Managed-competition health insurance in the Netherlands. N Eng J Med. 2007;357:2421–2423
    • View In Article
26. Edwards M. As good as it gets. AARP Magazine. 2004;47:42–4990
    • View In Article
27. Helton MR, Pathman DE. Caring for older patients: Current attitudes and future plans for family medicine residents. Fam Med. 2008;40:707–714
    • View In Article
28. Jepson C, Asch DA, Hershey JC, Ubel PA. In a mailed physician survey, questionnaire length had a threshold effect on response rate. J Clin Epidemiol. 2005;58:103–105
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (148 KB)
    • CrossRef
29. Robertson J, Walkom EJ, McGettigan P. Response rates and representativeness: A lottery incentive improves physician survey return rates. Pharmacoepidemiol Drug Saf. 2005;14:571–577
    • View In Article
    • MEDLINE
    • CrossRef