OUP user menu

Electronic health record use and preventive counseling for US children and adolescents

Cynthia M Rand , Aaron Blumkin , Peter G Szilagyi
DOI: http://dx.doi.org/10.1136/amiajnl-2013-002260 e152-e156 First published online: 1 February 2014

Abstract

The objective was to assess whether rates of preventive counseling delivered at well child visits (WCVs) differ for practices with basic, fully functional, or no electronic health record (EHR). Cross-sectional analyses of WCVs included in the National Ambulatory Medical Care Survey and National Hospital Ambulatory Medical Care Survey Electronic Medical Records Supplement, 2007–2010 were performed. Practices with fully functional EHRs documented delivery of 34% more counseling topics than those without an EHR. WCVs with a fully functional EHR lasted 3.5 more minutes than those with a basic EHR. Overall, for each additional 10 min, 12% more topics were covered, regardless of EHR functionality. Further studies should assess the impact of such systems on the quality of patient care during preventive care visits and on health outcomes.

Keywords
  • preventive health services
  • child health services
  • electronic health records
  • counseling
  • documentation

Introduction

Provision of anticipatory guidance to assist patients and their families in preventing multiple morbidities is a critical component of pediatric well-child visits (WCVs).1 Many professional organizations, including the American Academy of Pediatrics (AAP)2 and American Academy of Family Physicians (AAFP),35 recommend counseling all pediatric patients about diet and nutrition, exercise, injury prevention, and tobacco exposure, among other topics, at every WCV. Electronic health records (EHRs) can provide clinical decision support (CDS) to guide providers in discussing preventive topics. Use of CDS has been shown to improve the care of children with asthma6 and ADHD.7 While the use of EHRs in the USA is expanding with an influx of governmental support,8 most are initially designed with adult patients in mind, and the impact of EHRs on pediatric WCVs is unknown. Prior studies have assessed anticipatory guidance delivered at visits with paper compared to computerized data entry,9 but none have yet examined preventive services based on the functionality of an EHR. We analyzed a nationally representative sample to assess whether functionality of EHRs is associated with the length of WCVs and with reported rates of preventive counseling.

Methods

Data source

We performed cross-sectional analyses, combining data from the National Ambulatory Medical Care Survey (NAMCS) and National Hospital Ambulatory Medical Care Survey (NHAMCS) Electronic Medical Records Supplement from 2007–2010. Both surveys are collected annually by the National Center for Health Statistics (NCHS), and are publically available. NAMCS provides information about the use of ambulatory medical care services; NHAMCS provides details about hospital-based outpatient and emergency departments in the USA. Both use a multistage probability sampling design for physician practices and patient visits within practices, which, when weighted, provides nationally representative estimates of ambulatory visits in the USA. For each visit, the physician or a staff member completes a survey form that identifies patient demographics, counseling topics discussed, International Classification of Diseases, 9th edition (ICD-9) codes, and the visit duration.

Measures

We included WCVs (defined by ICD-9 codes V20.2, V70.0, V70.3, V70.5, V70.6, V70.8, V70.9) made to pediatric, family medicine, or generalist offices by children and adolescents aged 0–18 years.10 As defined by researchers at the NCHS,11 we categorized EHR use as none, basic, or fully functional, with the following criteria: basic systems have the ability to record patient history and demographics, problem lists, physician clinical notes, and medication lists, perform computerized prescription ordering, and allow lab and imaging result viewing; fully functional systems do all of these functions, and also include drug interaction warnings, ability to send prescriptions directly to pharmacies, computerized orders for lab tests, highlighted out-of-range lab values, medical history and follow-up notes, and reminders for guideline based interventions or screening tests. Providers were able to answer ‘yes’, ‘yes, but turned off or not used’, ‘no’, or ‘unknown’ to each function. Providers answering ‘yes’ to each item noted above were labeled as basic or fully functional users of the EHR. These items are included in the meaningful use ‘Core Set’ objectives. No EHR use was defined by participants who answered ‘no’ to the question ‘Does your OPD [outpatient department] use electronic medical records (not including billing records)?’

We delineated child age (infant, <2; child, 2 to <11; adolescent, 11–18); race/ethnicity (White, Black, Hispanic, or other); insurance (private, public, other); and percentage of poverty in the patient's zip code (<5%, 5 to <10%, 10 to <20%, and ≥20%). We identified the practice owner (physician, other hospital, or other); specialty (pediatric or general medicine); and provider type seen at the visit (Doctor of Medicine or Osteopathic Medicine (MD/DO), or physician assistant/nurse practitioner (PA/NP)).

We examined rates of counseling for the following topics (as defined by survey questions): diet/nutrition, exercise, growth/development, injury prevention, tobacco use/exposure, any health education (any of these listed topics), and other health education (any topic other than those specifically listed). We also assessed the total number of topics discussed, and the time spent at the visit (NAMCS only, not included in NHAMCS).

Analyses

We used χ2 tests to compare the differences for categorical demographic variables and topics being discussed between the three different EHR categories. We used negative binomial regression to test for differences between the EHR categories for the total number of topics being discussed, for both bivariate and multiple regression, since this was an over-dispersed discrete variable. All analyses were done using Stata V.12.1, and were weighted to be nationally representative, adjusted for the complex sampling design.

Results

A total of 8650 visits were included in the dataset, representing an annual average of 47 million visits nationally. Given the higher rate of recommended visits in the first year of life, almost half (45%) of visits were for children <2 years of age (table 1). There were no significant differences in EHR use based on patient or practice demographic characteristics. Overall, 77% of preventive visits were performed with no EHR, 14% with a basic EHR, and 9% with a fully functional EHR. Finally, 89% of WCVs had no EHR in 2007, compared to 65% in 2010; 3% were fully functional in 2007 compared to 17% in 2010.

View this table:
Table 1

Sample demographics and practice electronic health record (EHR) functionality

Visits (%) (n=8650)No EHR (%)Basic EHR (%)Fully functional EHR (%)p Value
Age (years)0.14
   <245454739
    2 to <1135353739
   11 to <1920201622
Sex
   Female484946470.70
   Male52515453
Race/ethnicity0.42
   White59596359
   Black13141111
   Hispanic21221822
   Other6589
Insurance0.19
   Private62635768
   Public34333630
   Other4471
Percent poverty0.48
   <5%29302530
   5 to <10%28273726
   10 to <20%28282827
   ≥20%15151017
Provider specialty0.38
   Pediatric82827784
   General medicine18182316
Dataset0.45
   NAMCS94959293
   NHAMCS6587
Practice owner0.77
   Physician74757270
   Other hospital13131213
   Other13121616
Year0.01
   20072428148
   200826253223
   200925252222
   201025223248
Provider type0.08
   MD98989599
   PA/NP251
   All (millions of visits)77 (36.2)14 (6.7)9 (4.3)
  • MD, Doctor of Medicine; NAMCS, National Ambulatory Medical Care Survey; NHAMCS, National Hospital Ambulatory Medical Care Survey; NP, nurse practitioner; PA, physician assistant.

The highest rates of counseling were for diet/nutrition (47%, 45%, and 57%; p=0.33), growth (46%, 45%, and 57%; p=0.37), and any health education (68%, 72%, and 75%; p=0.54) for none, basic, and fully functional EHRs, respectively (figure 1). While diet/nutrition, exercise, growth and injury counseling trended toward being higher at visits with a fully functional EHR, these differences were not significant.

Figure 1

Counseling delivered based on electronic health record (EHR) status*

Time (in minutes) spent at the preventive visit (19.8 vs 18.6 vs 22.1 min; p=0.12) for none, basic, and fully-functional EHR practices, respectively, were similar when examined overall. If we strictly compare basic to fully functional EHRs, the differences in time spent are significant, with fully functional EHR visits taking 3.5 min more (18% longer) than those with basic EHRs (p=0.05).

Table 2 shows the total number of topics discussed at the preventive visit. Controlling for patient age, insurance, level of poverty, and provider specialty, practices with a fully functional EHR documented delivery of 34% more counseling topics than those without an EHR (p=0.01). In addition, patients aged 2 to <11 years (p = 0.002) and 11 to ≤19 years (p=0.02) received 11% and 12% more counseling topics, respectively, than infants. Individuals living in areas with ≥20% poverty received 20% less counseling compared to those in areas with <5% poverty (p=0.04). General practitioners provided 43% fewer counseling topics compared to pediatricians (<0.001).

View this table:
Table 2

Factors associated with total number topics discussed at a preventive visit for all visits

Total topics IRR (95% CI)p Value
EHR practice
   None/partial EHR (Ref)1.0
   Basic EHR0.98 (0.74 to 1.28)0.87
   Fully-functional EHR1.34 (1.07 to 1.68)0.01
Age (years)
   <2 (Ref)1.0
   2 to <111.11 (1.04 to 1.19)0.002
   11 to <191.12 (1.02 to 1.24)0.02
Insurance
   Private (Ref)1.0
   Public0.98 (0.85 to 1.12)0.75
   Other0.92 (0.74 to 1.13)0.42
Percent poverty
   <5% (Ref)1.0
   5 to <10%0.98 (0.83 to 1.16)0.80
   10 to <20%0.90 (0.74 to 1.09)0.28
    ≥20%0.80 (0.65 to 0.99)0.04
Provider specialty
   Pediatrics (Ref)1.0
   General medicine0.57 (0.46 to 0.70)<0.001
  • EHR, electronic health record.

When time spent at the visit is included in the model, visits with fully functional EHRs provided 36% more counseling than those without an EHR (p=0.009). For each 10-min increase in the time spent, the average number of topics discussed increased by 12% (p=0.01), irrespective of EHR functionality.

Discussion

This is the first study to examine the quality of care delivered at pediatric preventive visits based on the functionality of a practice's EHR system. Based on a nationally representative survey, we found that visits performed at practices with fully functional EHRs documented delivery of 34% more counseling topics than those without an EHR. Notably, 18% more time was spent at WCVs with fully-functional compared to basic EHRs, and for each 10-min increase in time spent at the visit, 12% more topics were documented as being discussed.

Rates of EHR use in our study were similar to a 2008 national study of adult and pediatric ambulatory practices (13% basic, 4% fully functional),12 but lower than the rate of 32% in 2009 of ‘advanced stage’ EHRs reported for inpatient stays for children in 2009.13 With government stimulus money to support regional extension centers and bonuses to providers for using EHRs in a meaningful way built into the American Recovery and Reinvestment Act of 2009, the number of practices using ‘fully-functional’ EHRs will increase rapidly, as occurred between 2009–2010 during which fully-functional practices increased from 8–17%, and others have noted.14

The number of counseling topics reported as discussed was highest at visits with fully-functional EHRs. We must caution that (a) the database includes topics reported to be discussed, and (b) we cannot necessarily assume causation in this cross-sectional study. However, we speculate that this association with functionality of EHRs could be related to age-specific templates or cues incorporated into EHR templates; such structured WCV templates have been shown to improve provision of anticipatory guidance.15 ,16 This is promising as more practices move to using such systems. It is unclear what EHR component might have increased the counseling provided, since specific components of CDS were not measured in NAMCS/NHAMCS. In addition, the number of topics discussed do not equate to the quality of discussion or to patient behavior change. Future studies are needed to assess these issues.

Rates of counseling were not significantly higher for any specific topic, although there appears to be a trend toward greater discussion of each topic when a fully-functional EHR is used. Only diet and growth were discussed at half of visits; all other topics were discussed less frequently, with tobacco counseling noted at only 11% of visits overall. Of note, documenting tobacco usage and providing tobacco cessation counseling for patients over age 13 is a meaningful use criterion to receive reimbursement from the Center for Medicaid Services.17 Therefore it is likely that this measure will improve dramatically for practices using EHRs. In a 2011 Cochrane review,18 documentation of tobacco status did increase, as did referral to tobacco cessation counseling, when an EHR was in place. Documentation was moderate for diet counseling, but quite poor for exercise counseling. These topics are categorized as alternative quality measures for meaningful use, so it is critical for pediatric-oriented EHRs to integrate such documentation into the workflow for providers to both provide and receive credit for such counseling.

Studies vary on whether computerized systems add time to visits, with one study noting an additional 5 min spent at pediatric preventive visits,9 and others showing no change.19 ,20 We found that visits with a fully functional EHR took 3.5 min longer than those with a basic EHR, which could relate to the additional counseling provided. However, we do not have data about how long each practice has used its EHR, which may affect visit length. While the definition of fully functional EHRs did include provider prompts, we do not have details about the type, placement, or use of prompts which must be easily visible, timely, and integrated into the provider's workflow to be effective.21 ,22 Nevertheless, an increase of 18% in visit length for WCVs, if confirmed, represents a substantial lengthening of pediatric visits.

While the long-term goal of a fully-functional EHR is to improve the quality of care delivered, simply having such a system does not ensure improved quality.23 ,24 Studies have shown that EHR systems are often not used to their full capacity,25 but clinicians who have used structured documentation and have relied on CDS provided better care for specific chronic conditions,26 had improved screening,27 ,28 and delivered better preventive care.29 Ideally, a pediatric-oriented EHR will be able to link to immunization information systems, provide decision support for immunizations, and dose medications by body weight, among other functions.30 We are hopeful that the finding of increased counseling provided with a more functional EHR will carry over to other improved abilities to provide pediatric preventive care.

Limitations

The counseling provided and time spent at the visit was documented by the provider, and could result in over- or under-reporting.31 Of note, the WCV times in our study were similar to the average time of 20.3 min for WCVs noted in a recent study (observed and timed).32 One other study has shown that providers over-report time spent on the NAMCS survey33; however we would not expect under- or over-reporting to vary based on EHR status. Also, we have no information about what CDS was in place.

Conclusion

In this national study, we found that 34% more counseling topics were documented at pediatric preventive care visits in practices using a fully-functional EHR compared to those without an EHR, but visits using fully-functional EHRs lasted longer than WCVs using basic EHRs. Further studies should assess the impact of EHRs on patient care quality and outcomes.

Contributors

CMR participated in the conception and design, analysis and interpretation, drafted and revised the article and approved the final version. PGS participated in the conception and design, interpretation of the data, draft of the article, and approved the final version. AB participated in the acquisition of data and analysis, and approved the final version of the manuscript.

Funding

This work was supported by the Agency for Healthcare Research and Quality grant number K08 HS017951.

Competing interests

None.

Provenance and peer review

Not commissioned; externally peer reviewed.

References

View Abstract