Use of Standing Orders for Vaccination Among Pediatricians

We conducted an Internet and mail survey on several topics related to childhood vaccination in June 2017 to September 2017 among pediatricians in a sentinel network of members of the AAP who spent ≥50% of their time practicing primary care. The Colorado Multiple Institutional Review Board reviewed and approved the study as exempt research not requiring written informed consent.

We developed this survey as part of the Vaccine Policy Collaborative Initiative, a collaboration with the Centers for Disease Control and Prevention to assess physician attitudes and practices related to vaccination. We developed a national network of pediatricians by recruiting from the AAP membership list and conducted quota sampling to develop a sentinel network of physicians. By using demographic data from the AAP membership list, quotas were established for a representative sample with respect to region (Northeast, South, Midwest, and West), practice location (urban, suburban, and rural), and practice setting (private practice, hospital or clinic, and health maintenance organization).²⁰  A random sample of primary care physicians were selected from the AAP membership list and invited to participate as sentinel physicians for the survey network. After enough physicians had been recruited to meet described quota categories, additional respondents within those full categories were excluded to maintain representativeness of the network. We have shown previously that sentinel network physicians were similar to physicians sampled randomly from the American Medical Association physician database when comparing demographics, practice characteristics, and attitudes about vaccination.²⁰ 

Survey items were developed by incorporating barriers to use of standing orders identified in published literature and expert opinion, including from key stakeholders at the AAP. In the survey, standing orders were defined as “a written or verbal policy that persons other than a medical provider, such as a nurse or medical assistant, may consent and vaccinate a person without speaking with the physician or advanced care provider first,” and in such cases, “a vaccine may be given before or after a physician encounter or in the absence of a physician encounter altogether.” Response options for the item measuring use of standing orders were “yes, for all routinely recommended vaccines,” “yes, for some vaccines,” or “no.” We did not assess use of standing orders in greater detail or for individual vaccinations (such as influenza). We used 4-point scales to assess barriers to the use of standing orders (“major barrier” to “not at all a barrier”). A national advisory panel of pediatricians pretested the survey. Then the survey was piloted among 41 other pediatricians and modified on the basis of feedback. The survey instrument is presented in Supplemental Fig 2.

The survey was sent via Internet (link to Web-based survey e-mailed to participants; Verint, Melville, NY [http://www.verint.com]) or by mail on the basis of physician preference. Using methods from previous surveys, we sent the Internet group up to 9 e-mail reminders, followed by 2 reminders sent to nonrespondents by postal mail.²¹  We sent the mail group an initial mailing, followed by a reminder postcard and up to 2 additional reminders with replacement questionnaires. The mail protocol was patterned on the tailored design method by Dillman et al.²² 

We combined Internet and mail survey responses for analyses because studies have demonstrated that respondent attitudes are similar when obtained by either method.^22–24  We compared respondents with nonrespondents using Student’s t tests, Wilcoxon rank tests, χ² tests, and Fisher’s exact tests, as appropriate. We conducted a multivariable analysis with the dependent variable of using standing orders for any routinely recommended pediatric vaccination. For this analysis, using standing orders was defined as a response of “yes, for all routinely recommended vaccines,” or “yes, for some vaccines” to the survey item assessing use of standing orders. Independent variables assessed included physician and practice characteristics and responses to questions regarding barriers to use of standing orders.

Barrier items were coded as 0 (not at all a barrier), 1 (minor barrier), 2 (somewhat of a barrier) and 3 (major barrier). To reduce the number of barriers in the multivariable model, a principal components analysis with varimax rotation was performed. Two factors were retained by using the criteria of an eigenvalue ≥1. Three barriers grouped on 1 factor, which we called “physician responsibility” (Cronbach’s α = 0.81; Table 1). Four barriers grouped on another factor, which we called “concerns about office processes” (Cronbach’s α = 0.77). We generated mean scores of the items grouped on each factor and used these scores in the multivariable model. The barrier “in my state, there are restrictions that limit the use of standing orders” was left out of the principal component analysis because it did not clearly assess a personal concern or perception as the other barrier items did, and because we did not have the ability to verify whether responses reflected actual legal restrictions or physicians’ perceptions of such restrictions.

Characteristics significant at P ≤. 25 in bivariate analyses were tested in multivariable models by using backward elimination, resulting in retention of factors that were significant at P <. 05 in the final model. Risk ratios (RRs) were calculated because of the tendency of odds ratios to overestimate effect sizes when outcomes are common. Calculation of adjusted RRs was conducted by using a log-Poisson model with robust error estimation (SAS PROC GENMOD; SAS Institute, Inc, Cary, NC).

On the basis of published literature showing an association between organizational factors and use of standing orders,^16,17  we used Wilcoxon rank tests to compare scores for physician responsibility and concerns about office processes factors between respondents whose practices make decisions about vaccines independently and respondents whose practices make decisions at a larger system level. All analyses were performed by using SAS version 9.4 (SAS Institute, Inc).

The response rate was 79% (372 of 471). Respondents did not differ significantly from nonrespondents by sex, age, number of years in practice, practice setting, urban or rural location, region, or locus of decision-making about vaccines (independent versus larger system level) (Table 2).

Fifty-nine percent of respondents reported using standing orders in their office, including 36% who use standing orders for all routinely recommended vaccines and 23% who use standing orders for some vaccines.

The most commonly reported barriers to use of standing orders among pediatricians who did not use standing orders were concern that patients may mistakenly receive the wrong vaccine (68% identified as a major barrier or somewhat of a barrier), concern that patients may prefer to speak with the physician about a vaccine before receiving it (62%), and belief that it is important for the physician to be the person who recommends a vaccine to his or her patients (57%). These top 3 barriers remained the same among all respondents and were reported as a major barrier or somewhat of a barrier by 40%, 39%, and 34%, respectively. Responses to additional barrier items are presented in Fig 1.

In bivariate analyses, several practice characteristics (census location and locus of decision-making about vaccines) and barrier items were associated with use of standing orders for any routinely recommended vaccines (Table 1). In the multivariable model, the only factor retained at P <. 05 was the mean score of items contributing to the physician responsibility barrier factor (RR: 0.59 per point increase; 95% confidence interval: 0.53–0.66).

Wilcoxon rank tests revealed that physician responsibility scores were higher among respondents whose practices make independent decisions about vaccines than among respondents whose practices make larger system-level decisions (P =. 05). There was no association between decision-making level and concerns about office processes scores (P = .27).

In this nationally representative survey of pediatricians, a majority of pediatricians report use of standing orders for at least some vaccines; however, 41% still do not use standing orders for any routinely administered vaccines. Among respondents who did not use standing orders, the most commonly reported barriers were concern that patients may mistakenly receive the wrong vaccine, concern that patients may prefer to speak with a physician before receiving a vaccine, and belief that it is important for the physician to be the person who recommends a vaccine to their patients. In the principal component analysis, these 3 barriers grouped on 1 physician responsibility factor, and the physician responsibility score was the only factor that remained associated with use of standing orders in the multivariable model.

Many survey respondents identified concern that patients might receive the wrong vaccine and concern about removing physicians from the conversation with patients about vaccines as barriers to use of standing orders, and these physician responsibility barriers were associated with decreased use of standing orders. In a study of patients attending obstetrics and gynecology clinics, most patients surveyed were comfortable with the idea of using standing orders for human papillomavirus vaccination.²⁵  In that survey, more patients supported the idea of using standing orders for human papillomavirus vaccination in adults (79%) than in adolescents (43%), suggesting that patient and parent attitudes about use of standing orders for childhood immunizations may differ from attitudes about adult vaccination. That survey also showed that acceptability of standing orders was greater if patients were first told that a provider strongly recommended the vaccine. Physician concern that patients or parents want to speak with the physician before receiving a vaccine may be well founded in some cases but could possibly be mitigated by having another staff member communicate the provider’s strong recommendation for vaccination. Although some pediatricians surveyed indicated the concern that patients would prefer to speak with a physician as a barrier, our survey did not include parents, and so we could not directly assess parental acceptability of standing orders for immunization. Studies showing that standing orders are associated with increased immunization uptake in pediatric settings would imply that standing orders were acceptable to many parents in those studies because more children were vaccinated when standing orders were in place.^4–8  The acceptability of using standing orders for child and adolescent vaccines warrants additional study among parents and patients, noting that attitudes toward adult vaccination practices may not translate to the pediatric setting and that physician concern about parental acceptability may not accurately reflect perceptions among parents.

Working in a practice where decisions about vaccines are made on a system level was associated with use of standing orders in bivariate analyses. This finding is consistent with studies of adult vaccination that have revealed practice-level and organizational factors associated with use of standing orders.^16–18  Decision-making level was not associated with use of standing orders in our multivariable model; however, pediatricians reporting more independent decision-making about vaccines had higher physician responsibility barrier scores. Perhaps physicians who feel more personal responsibility about their role in vaccination are more likely to choose practice settings where they have more independent decision-making ability. Alternatively, knowing the level of decision-making about vaccines in the practice may influence the amount of personal responsibility that pediatricians feel about their role in vaccine delivery.

Physician concern about patients receiving the wrong vaccine if they did not personally assess eligibility for vaccination was another item within the physician responsibility factor that was associated with decreased use of standing orders. Although there is evidence for the use of standing orders to increase vaccination rates, published studies have not included systematic assessment of errors in vaccine administration under standing orders. Sixty-eight percent of pediatricians who did not use standing orders reported this concern about patients receiving the wrong vaccine as a major barrier” or somewhat of a barrier, compared with only 20% of pediatricians who did use standing orders. The significant difference in perceptions about this barrier suggests that pediatricians who do not use standing orders may overestimate the frequency of vaccination errors or that pediatricians who use standing orders have found ways to prevent such errors. Using the EMR to link standing orders for immunization with decision-support tools regarding immunization eligibility may be one way to address physician concern about the safety and accuracy of standing orders for immunization.

Few respondents (7% overall and 10% of those who did not use standing orders) reported that state restrictions limiting the use of standing orders was a major barrier or somewhat of a barrier to their use in practice. One review of state laws revealed significant variability regarding which types of nonphysician personnel are permitted to perform specific immunization delivery tasks (reviewing immunization status, prescribing vaccines, and administering vaccines independently or under delegated authority).²⁶  In many cases, state law is silent on a particular task for a particular type of health care professional, making implementation of standing orders potentially challenging for practices. Incorporating a more detailed assessment of local legal and regulatory factors may add to the understanding of pediatricians’ use of standing orders and associated barriers.

Pediatricians who did use standing orders reported relatively few barriers. The most common concern was that patients would prefer to speak with their physician about a vaccine before receiving it, and this was identified as a major barrier or somewhat of a barrier by 24% of respondents. Two other items were identified as a major barrier or somewhat of a barrier by 20%, and the remaining items were less frequently endorsed as barriers. In contrast, 7 items were identified as a major barrier or somewhat of a barrier by >20% of respondents who did not use standing orders, including 3 barriers endorsed by >50% of respondents. Further study, potentially by using qualitative methods, is needed to understand the reasons for these differences in barriers and to identify strategies used to overcome perceived barriers.

Uptake of routine childhood immunizations is high compared with uptake of adolescent, adult, and maternal immunizations. Perhaps pediatricians do not view standing orders as necessary for routine childhood vaccines and are less likely to use the strategy for that reason. With this hypothesis in mind, efforts to promote use of standing orders among pediatricians may be more successful by focusing on areas of low vaccine uptake such as influenza or adolescent immunizations.

This study had several limitations. Although we found no difference in demographic and practice characteristics between respondents and nonrespondents, they may have differed in their use of and barriers to use of standing orders. We did not ask pediatricians about the age distribution or the number of patients in their practice, the proportion of their patients who were unvaccinated, or their EMR usage. We surveyed pediatricians for their reported use of standing orders but did not assess actual clinical practice. Neither use of other practice-level strategies to improve immunization (such as reminder and recall) nor differences in the use of standing orders for different specific vaccines were assessed. In addition, we did not verify specific local legal and regulatory factors related to standing orders for vaccination, and so we were not able to robustly interpret responses to the barrier item “in my state, there are restrictions that limit the use of standing orders.” Finally, we did not assess vaccination rates, so we were unable to evaluate whether use of standing orders was associated with higher vaccination rates.

Standing orders are an evidence-based strategy to increase uptake of recommended immunizations, yet >40% of pediatricians surveyed do not use standing orders. More information is needed about the acceptability of standing orders for pediatric vaccinations among patients and parents, although studies revealing improvement in vaccination rates with use of standing orders in pediatric practices suggest that acceptability is not a substantial barrier to success when standing orders are implemented. Future studies to better understand current practices and promote the use of standing orders should address physicians’ concerns related to their personal responsibility for vaccination as well as organizational factors.

We thank Lynn Olson, PhD, and Karen O’Connor (Department of Research, AAP); Arlene Weissman, PhD; and the leaders of the AAP for collaborating in the establishment of the sentinel network in pediatrics. We also thank all pediatricians in the network for participating and responding to this survey.

AAP

American Academy of Pediatrics

EMR

electronic medical record

RR

risk ratio