Visit Length in Pediatric Primary Care: Lessons from a

1 downloads 0 Views 2MB Size Report
drive for efficient use of physician, nurse, clerk, and exam room resources have an important influence on patient access to health care services, as well.11. •. 14.
PRACTICE MANAGEMENT (

Visit Length in Pediatric Primary Care: Lessons from a Pilot Study Hasmet Uner, MD, MS,* Farnaz Ghazi Nezami,t Mehmet Bayram Yildirim, PhD,* Fang long Dong, PhD,I Julie Wellner, RN, BSN,§ and Douglas D. Bradham, DrPH, MA, MPH11 As business drivers create pressure to see more patients in a given period, th ere is no reliable guidance regarding the optimal allocation of resources in ambulatory visits. Many pediatric primary clinics set appointment lengths in increments of " five minutes." Defining the appointment lengths for potentially longer visits by arb itrary increments (e.g ., twice the time for an acute visit) is a common "experientia l" scheme. However, how much additional time is really needed if the patient is new to practice or when the

, • , • ' • , • • ,

visit is arranged for preventive services is unknown. Identifyi ng t he misallocation of clinic resources is fundamental because misallocation reduces access for patients and increases practice costs. In this study, using a time-motion approach, we examined t he characteristics of 372 visits in a pediatric primary care clinic.

• ,

*Assistant Professor, Director of Clinical Informatics, Department of Pediatrics, University of Kansas School of Medicine-Wichita, 3243 E. Murdock, Suite 402, Wichita, KS 67208; phone: 316-962-2114; fax : 316-962-7679; e-mail: [email protected]. tPhD Student, Industrial and Manufacturing Engineering, Wichita State University, Wichita, Kansas; e-mail: [email protected]. ;tAssociate Professor, Department of Industrial and Manufacturing Engineering, Wichita State University, Wichita, Kansas; e-mail: bayram. [email protected]. I Department of Preventive Medicine and Public Health, University of Ka nsas School o f MedicineWichita, Wichita, Kansas; e-mail: [email protected]. §Teaching Associate, Department of Pediatrics, University of Kansas School of Medicine-Wichita; Healthy Steps Specialist, PKU, Galactosemia, Hemoglobin Clinic Nurse, KU Wichita Pediatrics Clinic, Wichita, Kansas; e-mail: julie.wellner@wesleymc. com. I!Robert J Dole VAMC, and Professor of Public Health Economics, University of Kansas, School of MedicineWichita, Wichita, Kansas; e-mail: [email protected] . Copyright 0 2013 by Greenbranch Publishing LLC.

KEY WORDS: Visit le ngth; primary care; process evaluation; care delivery; data collection.

ore than one-third of visits for office-based physician care are to primary care doctors in general practice. 1 The proportion of primary care visits is likely higher in pediatric age groups because of the special role of h ealth maintenance visits (HMVs) in this population. 2 Duration of clinical visits has been studied for three decades. In these studies, researchers mostly examined visit length as an important determin ant of quality of care, physician or patient-satisfaction, and outcomes.3• 10 However, the length of appointment slots; their availability; and the drive for efficient use of physician, nurse, clerk, and exam room resources have an important influence on patient access to health care services, as well. 11• 14 As business drivers create pressure to see more patients in a given period, 15 there is no reliable guidance regarding the optimal allocation of resources in ambulatory visits. Although monitoring the changes in visit length over time using National Ambulatory Medical Care Surveys· (NAMCS) and evaluating the variation based on patient age and provider characteristics help, 16' 21 the utilization

M

of the evidence from these studies is limited particularly in clinic-based process evaluation undertaken to reduce inefficiencies. Consequently, "hands-on experience" and "trial and error" methods are typically used. We believe this study is the first to describe the clinician, nurse, and room times per physician times from the process evaluation perspective in a pediatric setting. Many pediatric primary clinics set appointment lengths in increments of "five minutes" (i.e., 10, 15, 20). Defining the appointment lengths for potentially longer visits by arbitrary increments (e.g., twice the time for an acute visit) is a common "experiential" scheme. However, how much additional time is really needed if the patient is new to practice or when the visit is arranged for preventive services is unpublished. Identifying the misallocation of clinic resources is fundamental to "bending the cost curve" in ambulatory care because misallocation reduces access for patients and increases practice costs. In this pilot study, we examine visit, patient, and provider characteristics of care processes and predictors of time actually spent in care activities.

www.greenbranch.com

1

800-933-3711

363

364

Medical Practice Management I May/June 2013

g:;;:$::I;~::::Z:=:z:=================:::l

METHODS

Study Settings and Data Collection Data collection was carried out between October and December 2010 using time-motion techniques in apediatric teaching clinic where 10 attending physicians, 23 pediatrics and internal medicine-pediatrics residents, and 3 midlevel providers serve mostly Medicaid patients. The existing scheduling system was designed to allocate 15-, 30-, or 45-minute slots sequentially. Data collection was completed while the clinic was using a paper-based medical record system. The information about all visits (planned and completed) was collected in blocked half-days. Halfdays were chosen to represent the weekdays and typical work schedules of providers. Providers and nurses documented the begin- and end-time of each activity/encounter of face-to-face care (FTFC) with patients that lasted longer than one minute. Providers and nurses also documented interpreter involvement, medical trainee participation, and patient room-in and room-out times. Observers ensured completeness of data collection. Information collected on paper forms was transferred to MS Access (Redmond, Washington) tables where "total" FTFC and waiting times per visit were calculated. Parents/patients and providers were provided a description of the objectives and the scope of the study. The approval of local institutional review board was obtained before data collection began.

Variable Definitions The dependent variables included: face-to-face provider times (i.e., attending, resident, nurse-practitioner/physician assistant), nurse times, exam room times, and patient/ parent waiting times in the lobby and examination room. The independent variables were: visit type, patient age group, provider type, interpreter and trainee involvements in the care, and whether the patient was seen by a primary provider (to whose panel the patient was assigned) or an available provider. Visit type was defined as "acute care," "health maintenance, new patient," "health maintenance, existing patient," and "follow-up." Follow-up visits included "any ER, hospital discharge or outpatient follow-ups," "weightcheck," and "chronic/acute care follow-up" visits. Patient ages were categorized based on the schema used in the coding of Evaluation and Management Services (Current Procedural Terminology). During the calculation of total FTFC times, overlapping times by different providers, such as faculty physician and trainee or nurse, were counted under the time of "main" provider (i.e., to whom the visit was scheduled). Nurse screening times were not included in room occupancy times because routine screening occurred in designated areas outside the exam room. Lobbywaiting time was defined as the period from the time the

check-in process was completed at the front desk until the time the patient was invited for screening. Initial in-room waiting was defined as the period between the patient's and main provider's first entry into the exam room. Other in-room waiting time consisted of aggregation of waits in the exam room between the FTFC activities after provider's first entry.

Analysis A generalized model analysis with underlying gamma distribution and log link parameter was conducted using SAS version 9.3 (Cary, North Carolina) to explore the relationships between dependent and independent variables. All tests were set at the significance level 0.05.

RESULTS Although we planned to collect information on about 605 clinical visits in 16 half-days, only 506 patients came for their clinic visit, reducing our available total sample. After removing the visits that included unusual circumstances, such as circumcision (6), sibling group (57), and lab-and shot-only (20) visits, 423 clinical visit records were reviewed for completeness. Any visit missing at least one timestamp for any FTFC (n =51, 12%) was excluded. Excluded visits were reviewed for potential bias using total visit length (i.e., the period between check-in and exam room-out times), patient age, and visit and provider types. No difference was detected between inclusion and exclusion groups. Fifteen- and 30-minute scheduled allocations were planned in most of the visits. The average length of appointment and total visit durations based on visit types are shown in Table l. Considering the relatively small sample and similar heterogeneity inherent in acute and follow-up visits, in the final analysis, we collapsed these visit types into one category because there was no statistically significant difference between the categories in appointment length scheduled in the system or in the times actually spent. Table 2 documents the variation in total provider FTFC times based on the times allotted in the system. Two acute visits, for which 20 and 45 minutes, respectively, were allocated at the time of scheduling, were excluded in this analysis. In the univariate analysis, the lengths of aggregated FTFC times were 16.8 ± 7.9 minutes for acute care and follow- up visits, 20.6 ± 8.4 minutes for established patients' HMVs, and 25.1 ± 8.2 minutes for new patients' HMVs. The differences in all pairs of visit types were statistically significant (p