Prospectively identifying patients at increased risk for hospitalization is a key prerequisite to designing interventions to prevent avoidable admissions. Primary care physicians (PCPs), with their comprehensive understanding of the clinical and psychosocial needs of their patients, may be uniquely qualified to identify risk factors missed by existing predictive algorithms. We examined whether PCPs could accurately identify patients at high risk for hospitalization in the subsequent year.

We invited PCPs in our network managing a panel of at least 100 patients in 2013 to review a list of 100 randomly selected patients from their panel.

 After verifying that the patient was theirs, PCPs were asked, “Would you be surprised if this patient was admitted to the hospital in the next year?” Patients were categorized as high or low admission risk based on whether PCPs answered “no” or “yes,” respectively.

We compared baseline demographics and Charlson Comorbidity Index (CCI) scores between patients in the PCP-designated high versus low admission risk groups using 

 tests, as appropriate. Unadjusted rates of hospitalization in the prior year and subsequent 1 year were compared using zero-inflated Poisson (ZIP) regression while accounting for clustering by PCP. To determine if PCP assessment remained an independent predictor of future hospitalization, we used ZIP regression to adjust for patient demographics, PCP characteristics, past acute care utilization, and CCI score, while accounting for clustering by PCP. We also calculated the C statistic for PCPs’ prediction of subsequent hospitalization.

Among eligible PCPs, 54% (99 of 182) reviewed their patient list. There were no statistically significant differences between participating and nonparticipating PCPs. Of 9900 eligible patients, PCPs reviewed 9832 patients and designated 238 (2.4%) as not being their patient. Among the remaining 9594 patients, PCPs designated 2037 (21.2%) as high admission risk and 7538 (78.6%) as low admission risk. Nineteen (0.2%) patients were unassigned and excluded from analyses.

Compared with patients in the low admission risk group, patients in the high admission risk group were significantly older, more likely to be white, insured by Medicare, lived in neighborhoods with a low median household income, had a higher Charlson Comorbidity Index score, and had increased unadjusted rates of acute care utilization in the past year and subsequent 1 year (Table).

PCPs’ qualitative assessment of future hospitalization risk among patients in their panels was an independent predictor of subsequent hospitalization, identifying a group of patients with a 172% increased rate of hospitalization (adjusted rate ratio, 2.72; 95% confidence interval [CI], 1.95-3.78). Predictive accuracy of PCP assessment in our study (C statistic, 0.77; 95% CI, 0.75-0.79) was comparable to the reported C statistic of other commonly used risk stratification instruments.

 There was no major difference in C statistic among those with (0.77; 95% CI, 0.72-0.81) or without (0.75; 95% CI, 0.73-0.77) hospitalization in the prior year, highlighting the fact that PCPs were not simply identifying those with past hospitalizations. Given the predictive accuracy of PCPs’ clinical assessment, efforts to identify patients at high risk for future hospitalization should aim to incorporate the unique insight that PCPs have about predisposing biopsychosocial factors. 

 Division of General Internal Medicine, Massachusetts General Hospital (ASH, JMA, WH, SJA), Boston MA; Los Angeles County Department of Health Services (CSH), Los Angeles, CA.

 Funded by Partners Community Healthcare, Inc, and by the Massachusetts General Hospital Primary Care Operations Improvement Program.

 The authors report no other relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

 Concept and design (JMA, SJA, ASH, CSH); acquisition of data (JMA, SJA, CSH, WH); analysis and interpretation of data (JMA, ASH, WH); drafting of the manuscript (ASH); critical revision of the manuscript for important intellectual content (JMA, SJA, ASH, CSH); statistical analysis (ASH); provision of patients or study materials (WH); obtaining funding (CSH); administrative, technical, or logistic support (CSH); and supervision (SJA, CSH).

 Andrew S. Hwang, MD, MPH, Gray Bigelow 730, 55 Fruit St, Department of Medicine, Boston, MA 02114. E-mail: ahwang1@partners.org.

1. Atlas SJ, Chang Y, Lasko TA, Chueh HC, Grant RW, Barry MJ. Is this “my” patient? development and validation of a predictive model to link patients to primary care providers. 

2. Haas LR, Takahashi PY, Shah ND, et al. Risk-stratification methods for identifying patients for care coordination. 

Articles

WEB EXCLUSIVE

This study demonstrates that the predictive accuracy of primary care physicians’ assessment of future hospitalization risk is comparable to commonly used quantitative risk stratification instruments.

Can Primary Care Physicians Accurately Predict the Likelihood of Hospitalization in Their Patients?

An Asthma Medication Ratio (AMR) of ≥0.50 has gained recognition as a metric related to improved asthma outcomes. The goals of this study were to evaluate asthma outcomes in relation to this threshold in Hispanic children with high-risk asthma and to compare the effectiveness of the Breathmobile (BM) program with usual care (UC), utilizing the AMR-related emergency department utilization (ED) rate.

 Healthcare utilization and prescription claims recorded January 1, 2011, to June 23, 2014, were evaluated pre- versus post year in 164 Medicaid-enrolled children, aged 2 to 18 years, with high-risk persistent asthma (BM group: n = 72; UC group: n = 92).

High risk was defined by ≥2 oral corticosteroid fills, or ≥2 ED visits (

International Classification of Diseases, Ninth Revision, Clinical Modification

 codes 493.0-493.9) in the pre-year. Poisson regression was utilized to compare asthma outcomes pre- versus post year in BM and UC groups and to describe the average number of ED days per 100 patients treated in relation to AMR.

 An AMR of ≥0.50 versus <0.50 was related to 49% fewer ED visits (

 <.05). BM proved to be more effective than UC in increasing the percent who achieved an AMR ≥0.50 from a baseline of 47.2% to 80.6% versus 50.0% to 65.2% post year, respectively. This difference was reflected in the significant 52% reduction in average number of ED visits post year for the BM cohort versus 13% for the UC cohort.

 View achievement of the AMR ≥0.50 as an informative metric in program evaluation and for healthcare organizations to measure the quality of care provided to high-risk patients with asthma.

An asthma medication ratio (AMR) of at least 0.50 reflects emergency department (ED) utilization and demonstrates superiority of using the Breathmobile (BM) compared with usual care (UC) in Hispanic children with high-risk persistent asthma.

		The AMR is an effective metric for assessing quality of care provided to asthmatic patients.

		Patients achieving an AMR ≥0.50 versus <0.50 had 49% fewer ED visits, 

		BM proved to be more effective than UC in increasing the percent of patients who achieved ≥0.50 from a baseline of 47.2% to 80.6% versus 50.0% to 65.2% post year, respectively.

		This AMR difference was reflected in the 52% reduction in ED visits post year in the BM cohort compared with 13% in the UC cohort.

Asthma, one of the most common chronic diseases in children, is estimated to affect more than 10 million children younger than 18 years in the United States.

 In those younger than 15 years, asthma exacerbations result in approximately 640,000 emergency department (ED) visits and 128,000 inpatient (IP) stays annually.

 Lifetime total direct medical costs of asthma per annual birth cohort are estimated at $3.2 billion.

 In addition, asthma is the leading cause of school absenteeism.

 African American and Hispanic children are disproportionately affected by relatively high morbidity rates.

 Morbidity is particularly common among those living in poverty, with poor access to medical care, and in those receiving public insurance.

 For example, Medicaid-covered children typically have poorer asthma control, greater underutilization of long-term controller medication compared with those receiving commercial insurance, and more frequent ED visits and IP stays.

As a result of these poor outcomes, efforts have been made to determine quality-of-care metrics that could predict future exacerbations in order to improve asthma management. These efforts led the National Committee for Quality Assurance to develop the Use of the Appropriate Medication for People with Asthma Metric as part of the Health Effectiveness and Data Information Set (HEDIS) to measure the quality of care provided to patients with asthma by healthcare organizations.

 There are 3 HEDIS measures that assess asthma care for those with persistent asthma.

 The initial asthma measure, reported in the year 2000, assessed the percentage of patients with persistent asthma who were prescribed at least 1 canister of controller medication in the measurement year. This idea has been examined by several studies, which failed to demonstrate a correlation of this measure with improved asthma outcomes.

 A second asthma HEDIS measure, the Medication Management for People with Asthma (MMA), first publicly reported in 2013, calculates the percentage of asthma controller medication adherence rate per treatment period and was reported in a recent study to not relate to improved asthma outcomes.

 The third HEDIS asthma measure, the Asthma Medication Ratio (AMR), has been shown to correlate with improved asthma outcomes in children and adults,

 and is currently being used as a quality-of-care metric to assess healthcare programs. However, potential variability in the AMR related to age, ethnicity/race, asthma severity, and type of insurance suggested to us the need for further evaluation in our population of Medicaid-insured Hispanic children with high-risk persistent asthma.

A program utilizing mobile clinics, the Breathmobile (BM) program, has previously demonstrated improvement in asthma control and reduction in healthcare utilization by providing access to preventative asthma care to underserved children.

 However, these studies had limited ability to examine improvement relative to a comparator group and by utilization of claims data.

We were fortunate to collaborate with the Medicaid program in Orange County, California, called CalOptima. They gave us the opportunity to access encounter and pharmacy claims data in order to evaluate asthma morbidity, use of oral corticosteroids (OCS), and AMR scores. In addition, we were able to compare those receiving BM care with a usual care (UC) group. The first goal of this study was to evaluate the AMR as a measure for determining improved asthma outcomes. The second goal was to evaluate the effectiveness of the BM healthcare program by comparing BM care with UC using the AMR quality improvement measure, as well as determining healthcare utilization rates in our population of Hispanic children with high-risk persistent asthma.

CalOptima generated an annual list of insured children eligible for participation in the BM program. Parents of eligible children were contacted and offered care. Those who agreed to participate were designated the BM cohort; those who did not, the UC cohort. The reasons for not enrolling in the BM cohort included: unable to contact family (>50%), family preference, those who were scheduled to be seen but did not show up to their appointment, and children who had received care with an asthma specialist in the previous 12 months. All patients were offered the CalOptima disease management intervention prior to referral to the BM program. The intervention included reminder letters for tests and recommended immunizations, quarterly screening and monitoring by a disease management coordinator, quarterly educational paper mailings reviewing self-management techniques, and a list of self-management classes. Patients were referred for care in the BM program from July 2011 to April 2013. Eligibility criteria specific to this study included: being 18 years or younger, diagnosis of asthma (

] codes 493.0-493.9) documented in encounter data, and being considered high risk. The criteria for being considered high risk are: persistent asthma as defined by HEDIS criteria

 and validated in the article by Schatz et al,

 ≥2 courses of OCS, and/or ≥2 ED visits (

 codes 493.0-493.9), and/or ≥1 IP stay (

 codes 493.0-493.9) in the previous 12 months. Other criteria included Hispanic ethnicity and a measurable AMR over the study period. Continuous CalOptima coverage during the entire study period was also required.

Encounter and prescription claims recorded from January 1, 2011, to June 23, 2014, in an electronic health record system utilized by CalOptima were abstracted for 164 children who met study eligibility criteria (n = 72 BM group and n = 92 UC group). Encounter claims data included age, gender, race, date of service, location of visit, diagnosis code(s), Current Procedural Terminology code, place of service, and whether patient had continuous insurance coverage. Prescription (Rx) claims data included date filled, label name, drug type, and days supplied. National Drug Codes identifying asthma medications and the label name specified in the claim line were evaluated to qualify medication prescribed (Rx) as a reliever, controller, or oral corticosteroid for the purposes of calculating HEDIS AMR ([number of controllers Rx] ÷ [number of controllers Rx + number of relievers Rx]) during respective annual evaluation periods. Claims data were aggregated on an exposure period basis (pre- and post year) to determine total number of defined events. Exposure time for patients in the BM group was determined from date of initial BM visit. In the UC group, pre- and postyear claims data were provided by CalOptima. This protocol was reviewed and approved by the Children’s Hospital of Orange County Institutional Review Board.

Study population characteristics were described in terms of valid percent of patients with a defined trait and the significance of between-intervention group (UC vs BM) differences assessed by 

AMR was calculated in pre- and postyear periods and described in terms of average value and percentage of members with an AMR ≥0.50. Increased percentage of patients with AMR ≥0.50 in the post compared with the pre-period was assessed for significance using McNemar’s test. Correspondence of AMR to ED visits, IP stays, and OCS fills was evaluated using Poisson regression with adjustment for patient age, gender, intervention group, and time period. Inclusion of an interaction effect between group and time determined significance of intervention group differentials in average medication fills. In SAS version 9.2 (SAS Institute Inc, Cary, North Carolina), PROC GENMOD with repeated statement was run to produce robust standard errors for the Poisson regression coefficients, controlling for potential overdispersion. In terms of reduced healthcare utilization, intervention group differentials were nonsignificant, but the small number of events (particularly IP stays) and sample size were factors impacting power to detect significance of interaction effect (results not shown). Calculation of incidence rate ratios (IRRs) were offset by patient-level exposure time for outcomes investigated per person-year in the BM group (pre- vs post period: retrospective look-back up to 365 days from initial BM visit vs elapsed time from initial BM visit to 365 days later or to cohort data end point [whichever occurred first]). Poisson regression analyses were performed using SAS version 9.2 (SAS Institute Inc, Cary, North Carolina) while remaining analyses utilized SPSS version 18.0 (SPSS Inc, Chicago, Illinois).

In the BM group, 65.3% were male and the average age was 7.1 years with 27.8% under 5 years, 54.2% aged 5 to 11 years, and 18.1% aged 12 to 18 years (

). The UC group was similar in age and gender composition; however, the percentage of patients in the BM group who required an ED visit (pre-year) was slightly higher than observed in the UC group (37.5% vs 23.9%; 

= .059). Prior to enrollment in the BM program, 47.2% of high-risk patients met HEDIS criteria (AMR ≥0.50). The percentage was similar in the UC group (50.0%), 

=.724. Nearly all patients in both groups reported OCS fills pre-year (98.6% in BM and 98.9% in UC group) and most had been prescribed at least 1 controller medication (86.1% and 80.4%, respectively). In terms of engagement in the BM program, 74% of patients had at least 3 visits to the program with a median time from baseline to third visit of 71 days (interquartile range: 57-107 days) by the study endpoint.

Achieving an AMR ≥0.50 corresponded to 49% fewer ED visits per 100 patients treated, after adjustment for age, gender, intervention group, and time period (

). Average number of ED days per 100 patients ranged from 50 in high-risk patients not using any controller medication (AMR = 0.00) down to an estimated 19 days in those using controller medication(s) without the requirement of relievers (AMR = 1.00). Each 0.10 increase in AMR resulted in 9% fewer ED days (

= .059). AMR did not show a statistically significant relationship to OCS fills and IP stays during the concurrent period (

Reduced Healthcare Utilization and Improved Pharmacological Outcomes

Among children whose parents accessed care offered through CalOptima in the CHOC Health Alliance BM, reduced average number of total claims (non-Rx), ED visits, IP stays, oral corticosteroid prescriptions, and increased controller medication use in the post compared with pre-period were significant (

). Positive trends were observed in the UC group, but to a lesser extent, with the exception of reduction in patients requiring an IP stay. The BM group had 52% fewer ED days per 100 patients treated post year (

= .050) compared with only 13% fewer in the UC group (

= .662). Overall, the total claims (non-Rx) rate showed a greater reduction in the BM compared with UC cohort ([IRR, 0.59; [standard error (SE) = 0.08] versus IRR, 0.65 [SE = 0.07], respectively). In our population of patients who met the criteria for high-risk disease (baseline), only 22.6% of BM patients continued to have high-risk disease post year. In the UC cohort, improvement was also observed, although the percentage remaining at risk was slightly higher at 32.6%.

The percentage of patients who met HEDIS AMR ≥0.50 increased from 47.2% to 80.6% in the BM group (

<.001) and from 50.0% to 65.2% in the UC group (

). Although BM and UC groups were similar in their average AMR pre-year (0.42 [standard deviation (SD) = 0.24] and 0.40 [SD = 0.25], respectively; 

= .649), the BM group had significantly higher average AMR post year compared with the UC group (0.60 [SD = 0.22] and 0.47 [SD = 0.28], respectively; 

=.002). A significant shift in the medication usage pattern helped approach the benchmark of OCS Rx <1 per year in the BM group (pre vs post: 2.6 vs 1.0; 

<.001) (Table 2). Trends were directionally similar in the UC group (2.9 vs 1.4; 

<.001). This translated to 60% fewer OCS fills per 100 patients treated in the BM group compared with 52% fewer in the UC group. Increased controller medication fills in the BM group was in line with expectations post year for treating individuals with high-risk disease (7.6 compared with 3.9 pre-year; 

<.001), whereas, a slight decrease in controller medication fills was observed in the UC group from 3.9 to 3.5 per year (

= .164). The difference in pharmacological therapy approach between the BM and UC groups was significant (

In this report, we have demonstrated the AMR to be an effective metric by assessing its relationship to improved asthma outcomes, particularly reduced ED visits in Hispanic children with high-risk persistent asthma. Utilizing the established optimal AMR threshold of ≥0.50, there were 49% fewer ED days per 100 patients compared with those with an AMR <0.50 (

<.05). Furthermore, using the change in the AMR as a quality improvement metric found that BM increased the percentage with an AMR ≥0.50 from 47.2% to 80.6% in the post year observation, whereas the comparator, UC, showed an increase from 50.0% to 65.2%. This difference was reflected in the significant 52% reduction in average number of ED visits for the BM cohort versus 13% for the UC cohort. The impact of the BM program was also shown by the reduction of OCS courses in the post year to approximately 1 compared with 1.4 in the UC group, suggesting continued risk of poor asthma control in the latter.

The concept wherein a provider proactively monitors their patient’s health status outside of the office setting, particularly electronically, enables them to interact before severe asthma exacerbation or loss of asthma control occurs. This is particularly important with a disease like asthma where signs and symptoms may be episodic and healthcare utilization high. For that reason, administrative data are used in HEDIS and by the National Committee on Quality Assurance to assess the quality of care by health plans, such as Medicaid, and are increasingly incorporated into the pay-for-performance reimbursement model, which encourages a preventative approach to asthma therapy.

Of the 3 HEDIS process measures for persistent asthma, the AMR has shown superiority to ≥1 controller medication prescriptions per year.

 This was also evident in our population, as ≥80% of high-risk children had received ≥1 controller in the pre-year. Yoon et al recently evaluated the MMA as a quality-of-care metric and found the MMA was not related to improved asthma outcomes assessed by rescue medication prescriptions, ED visits, and IP stays.

One potential reason given for the superiority of the AMR, compared with the first 2 HEDIS process measures, is that they lack measure of short acting beta-agonist (SABA) prescriptions, which may indicate asthma exacerbations and poor asthma control. Schatz et al found the 2 measures that correlated with asthma outcomes were the number of SABA prescriptions and the AMR

—the latter includes both controller and rescue medication. These authors reported that the number of SABA prescriptions per year was inversely correlated with asthma outcomes. We hypothesized that the lack of change in SABA use in our BM cohort, which was less than 4 prescriptions, a threshold associated with increased risk of ED visits and OCS use,

 may reflect general recommendations for use as a safety precaution prior to vigorous exercise (eg, running or basketball).

In the pediatric population, particularly in those with persistent asthma, several studies have shown the AMR to be a useful quality-of-care metric in relationship to improved asthma outcomes. Rust et al evaluated a Medicaid population of children aged 5 to 12 years with either 1 hospitalization or 2 outpatient visits for asthma and found that only 33.4% had an AMR ≥0.50 over a 90-day follow-up period. Those with an AMR ≥0.50 had a 17% reduced odds of a future ED visit.

 The positive impact of achieving AMR ≥0.50 was also found in our cohort with the 49% reduction in average number of concurrent ED visits. Rust et al noted that the proportion of prescribed days (≥50% vs <50%), similar to the MMA, did not correspond to improved asthma outcomes,

 which mirrored the findings of Yoon et al.

 In a subsequent paper by Rust et al, utilizing the same population, these investigators established the cost savings with a greater percentage reaching ≥0.50 AMR as a marker of greater adherence.

 For example, they projected a cost savings of $523.53 per patient due to reduced healthcare utilization by increasing the percentage reaching this threshold from 33.5% to 70%. This 36.5% absolute percentage point increase is similar to the 33.4% improvement (from 47.2% to 80.6%) observed in our BM cohort. Prior cost-benefit evaluations of the BM program showed similar positive cost reductions but preceded standardization on achieving AMR ≥0.50 and evaluation of attributable cost savings.

 In a novel approach, Beck et al calculated a pharmacy-level AMR (PH-AMR) and found children in census tracts with a PH-AMR reaching the ≥0.50 threshold had significantly less utilization than those with PH-AMR <0.50 (

They reported that for every 0.1 increase in the PH-AMR, the asthma emergent care utilization rate decreased by 9.5 events per 1000 children (

= .03). After adjusting for poverty level and access to care, they concluded that the pharmacy may be a community leverage point to improve population-level asthma control through targeted interventions. For every 0.1 increase in AMR in our community setting, the rate of ED visits per 100 high-risk patients decreased approximately 9% (

Stanford et al found that the optimal AMR threshold, depending on population, type of insurance, and timeframe being evaluated, ranged from ≥0.50 to ≥0.70.

 However, AMR defined at commonly used and effective thresholds of ≥0.50 and ≥0.70 was found to be a significant predictor of subsequent exacerbations and OCS usage in both children and adults. This relationship was also seen in those receiving either Medicaid or commercial insurance, particularly in those defined as being high-risk for persistent asthma.

The concept of providing mobile healthcare to patients who have poor accessibility to adequate medical care is not new and has been successful in several rural and low socioeconomic status (SES)/underserved communities for both adults and children.

 The concept of the BM model of healthcare delivery was originated in Los Angeles by Jones et al in 1995.

 This program addressed several major barriers to preventative care including accessibility, cultural compatibility, affordability, and continuity of care. Outcomes shown by several BM programs distributed in underserved communities throughout the United States have consistently shown self-reported reduction in healthcare utilization, improvement in asthma control, and reduced school absenteeism.

Most studies evaluating the AMR as a population-based healthcare metric have utilized extensive Medicaid databases, while our study evaluated far fewer patients selected on the basis of being high risk for asthma exacerbations. This limitation may have affected our power to detect the significance of clinically meaningful differentials in health resource use reductions between intervention groups in the post period.

Restricting our study to low-SES Hispanic patients may have limited the generalizability of our findings. Although our study was not a randomized clinical trial, selection bias was expected to be minimal as both BM and UC patients met criteria for high-risk asthma and were similar in distribution of baseline characteristics described in Table 1. Access to disease management strategies offered through CalOptima may have contributed to the higher-than-expected baseline percentage of patients whose AMR value was ≥0.50 in both BM and UC cohorts (pre-year: 47.2% and 50.0%, respectively), based on the much lower percentage reported for a similar population in the Rust article (33.4%).

 Our approach differed from some other AMR studies in that it was not designed sequentially to predict changes in the postintervention observational period,

 but rather concomitantly to determine the relationship of AMR to healthcare utilization and pharmacological outcomes.

 Outcome improvement may occur during subsequent years as shown by others.

 Population health metrics, such as the AMR, provide insightful and comparative data for application in evaluation of disease management programs; approaches to increase accessibility to parameters necessary to calculate this measure across multiple settings and populations should be explored. 

We have demonstrated that the AMR is a useful quality-of-care metric because it is significantly related to improved asthma outcomes. In our study, those with an AMR ≥.50 had 49% fewer ED visits compared with those with an AMR <.50 (

<.05) in 164 high-risk Medicaid-insured Hispanic children. Our data suggest that the BM model of care was superior to UC by increasing the percent of patients who achieved an AMR ≥0.50 from a baseline of 47.2% to 80.6% versus 50.0% to 65.2% post year, respectively. This was associated with a reduction of 52% in ED visits in the BM program (

<.05) compared with 13% in the UC cohort (

≥.05). Although the AMR has been applied to assess large healthcare organizations, we have shown in a smaller cohort that it is an effective metric on quality of asthma care. Our findings should lead others to view achievement of the AMR ≥0.50 as an informative metric in program evaluation and for healthcare organizations to measure the quality of care provided to patients with asthma.

The authors thank Cristina Bernal for her assistance with manuscript preparation.

 Morphew Consulting, LLC (TM), Bothell, WA; Children’s Hospital of Orange County (WA, SPG), Orange, CA; University of California, Irvine (SLB, SPG), Irvine, CA.

 Ms Morphew has a consulting agreement with Children’s Hospital of Orange County, which funded the analysis. The remaining authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

 Concept and design (TM, WA, SLB, SPG); acquisition of data (TM, WA, SLB, SPG); analysis and interpretation of data (TM, WA, SPG); drafting of the manuscript (TM, WA, SLB, SPG); critical revision of the manuscript for important intellectual content (TM, WA, SLB, SPG); statistical analysis (TM); provision of patients or study materials (WA, SPG); obtaining funding (WA, SPG); administrative, technical, or logistic support (SPG); and supervision (WA, SPG).

 Tricia Morphew, MSc, Morphew Consulting, LLC, 2208 242nd St SW, Bothell, WA 98021. E-mail: tricia@morphewconsulting.com. 

1. Bloom B, Jones LI, Freeman G. Summary health statistics for U.S. children: National Health Interview Survey, 2012. 

2. Asthma & children fact sheet. American Lung Association website. http://www.lung.org/lung-health-and-diseases/lung-disease-lookup/asthma/learn-about-asthma/asthma-children-facts-sheet.html. Published September 2014. Accessed September 29, 2015.

An examination of the asthma medication ratio (â‰¥0.50) as an informative metric in program evaluation and for healthcare organizations to measure quality of care provided to patients with asthma.

The Breathmobile Improves the Asthma Medication Ratio and Decreases Emergency Department Utilization

The laboratory, as a diagnostic department in the hospital, plays an important role in the treatment and prevention of diseases. Paying attention to patients’ preferences for communication of test results may provide a better and more responsive system for delivering these results. This study aimed to identify patient preferences regarding receiving their laboratory test results electronically and to identify the reasons behind their choice.

This study was carried out in 2015 with 200 patients who had access to the internet and had been referred at least once previously to the hospital laboratory department to receive their test results. Data were collected through an expert-validated questionnaire, and its reliability was confirmed by test—retest (

 and marginal independence SPSS and R software.

 Ninety-eight percent of participants preferred to be notified by short message service when their test results were ready. All participants preferred to receive their test results online, and 82.5% (n = 165) preferred to receive both normal and abnormal test results this way. The main reason for receiving results online was time savings, which was reported by 77% of participants, followed by lowering the chance of missing the results (31%). About 40% of participants thought e-mail notification was more secure than accessing the results through a hospital website.

 Findings showed that although patients wanted to benefit from online services for receiving their test results, they were concerned about confidentiality and security. Before using online technologies, security measures necessary to protect patient privacy and to gain the trust of patients should be defined.

		Several problems in the communication of laboratory test results, such as prolonged wait time, erroneous results reported, and missing results, result in patient dissatisfaction. 

		Online services were introduced as a way to improve the accuracy and timeliness of information communication; patients also show interest in using online technology to communicate with their healthcare providers.

		We asked those patients with internet access about their preferences regarding the possibility of receiving their laboratory test results electronically. We found that they prefer receiving online results if the confidentiality and security of their information are ensured. 

		We recommend implementing the security measures necessary to protect patient privacy before deploying online patient—provider communication technologies.

The main goal of healthcare systems is to improve the public's health, which can be attained by the delivery of appropriate and required health services. Evaluation of a health system can be done through an assessment of its services.

 Obtaining patients’ opinions is a valid method for evaluating health services because patients are receiving healthcare and are a potential source of data; additionally, it is their right to have their opinions taken into consideration during processes of planning and evaluating.

 Patient satisfaction with a health institution’s services is among the most important tools for evaluation of system quality

; it can help staff and management to potentially revise their care methods.

The laboratory, as a diagnostic department of a hospital, plays a significant role in the treatment and prevention of diseases. Heeding patients’ opinions about lab services is important for improving provider—patient communication and enhancing the department’s performance. Laboratory Information System, applied in this department, is a computer-based system that manages data on clinical chemistry, hematology, bacteriology, etc. The components of this system include hardware, software, data, and working processes that allow the professionals who are responsible for data collecting and processing and distributing test results to do so efficiently.

Study results have shown that despite the existence of such systems in laboratories, patient dissatisfaction continues due to lack of system standardization and the systems’ poor performance.

 The most dissatisfaction relates to delays in receiving test results, missing results, or misfiling/misreporting the results.

 According to the findings of other studies, patients have great interest in using internet-based technology to communicate with their healthcare providers, to access their clinical records, and to obtain timely health information.

Using internet-based technologies to receive test results contributes to decreasing the wait time, increasing patient satisfaction, reducing multiple visits to receive the results, and, to some degree, possibly decreasing urban traffic. Moreover, such a system can have advantages for hospitals, such as requiring less laboratory/administrative staff to deal with paperwork, freeing up waiting rooms for other purposes, and saving the costs of printing lab results. However, the lack of face-to-face interaction with laboratory staff, and of their personal encouragement, may result in the failure of patients with abnormal lab results to plan for their next treatment steps. Still, a previous study showed that more than 80% of patients prefer to receive lab results electronically, such as via the internet.

In Iran, except in a few laboratories, online sending of test results is not widespread. Most patients still have to be present at laboratory departments to receive their test results.

To our knowledge, no study has addressed the preferences of Iranian patients regarding online delivery of laboratory test results. Our objective was to examine patient preferences, priorities, and recommendations for delivering laboratory test results electronically. This was a pre-implementation study conducted for the purpose of developing an online system for the delivery of test results. The findings can be of value for other organizations with similar goals as well.

This was a cross-sectional descriptive and analytic study carried out with the cooperation of patients visiting the laboratory department of a university hospital in Ferdows, Iran. A wide range of tests, including biochemistry, hematology, parasitology, bacteriology, serology, immunology, hormonology, pathology, and gasometry, is done in this department.

Patients were included in this study if they had 1) a previous history of receiving test results from this department, and 2) access to the internet. A sample of 195 patients was calculated using Cronbach’s alpha sampling method; ultimately, 200 patients were invited to participate.

A researcher-made data-collecting questionnaire was designed based on the review of literature.

 The validity of the questionnaire was confirmed by 4 medical informatics experts who have a background in medicine and health information management.

Its reliability was examined using the test-retest method (

The questionnaire consisted of 4 sections: 1) demographic data such as age, education, marital status, occupation, and residential address; 2) 2 questions about the time patients had spent to receive their test results in the previous encounter; 3) 12 questions about their use of information technology; and 4) 9 questions about the patients’ preferences and priorities for receiving test results, including their reasons, and their preferred method of test-results delivery (by e-mail; on the patient portal of a website, with or without short message reminder concerning presence of the results on the portal). One of the authors distributed the questionnaire to each patient visiting the lab who met the qualifications and then collected them after completion. The patients were instructed to complete the questionnaire, and their consent was inferred from their agreement to fill out the questionnaire.

Data analysis was done by SPSS software version 19 (IBM Corporation; Armonk, New York) and R software version 3.2.2 (R Foundation for Statistical Computing; Vienna, Austria) using descriptive and analytic statistics, including 

Data analysis showed that about half of the participants (n = 98; 49%) were 30 years or younger. Less participants had a bachelor’s degree (n = 78; 39%), most were married (n = 156; 78%), and almost half were employed (n = 88; 44%). The residential location of most participants (n = 131; 65.5%) was “city” (

In their previous encounter getting laboratory test results, fewer than half of participants (n = 83; 41.5%) had spent less than half an hour; 35.5% (n = 71) spent between 30 and 60 minutes; 19% (n = 38), between 60 and 120 minutes; and 4% (n = 8), more than 2 hours to reach the hospital laboratory. In addition, 42% of participants (n = 84) said that once they arrived at the hospital, they had to spend 5 to 15 minutes waiting for their test results. The wait time for 32.5% (n = 65) of participants was about 15 to 30 minutes; for 12.5 % (n = 25), more than 30 minutes. The rest (13%; n = 26) had to wait for less than 5 minutes.

Participants’ Internet Use and Their Access to the Infrastructure

In this study, all participants had access to the internet. Almost half (n = 94; 47%) mentioned that they use the internet less than 30 minutes daily, while 8% (n = 16) of participants used the internet more than 3 hours a day (

Most participants (65.5%; n = 131) used the internet often to access social networks (

). Most accessed the internet from home (n = 171; 85%), and 35% (n = 70) accessed it from work; 22.5% (n = 45) from public internet centers, 21.5% (n = 43) from friends’ homes, and 7% (n = 14) from their universities.

Of the participants, 74% (n = 148) had a personal e-mail address. The number of times they checked it included daily (n = 51; 34.5%), 1 to 2 times a week (n = 31; 20.9%), 3 to 4 times a week (n = 28; 18.9%), rarely (ie, less than once in a month) (n = 22; 14.9%), and 1 to 2 times a month (n = 16; 10.8%).

Forty-seven percent (n = 94) of the participants mentioned that they had already visited the hospital website. Among those, 84% (n = 79) reviewed the website to check the working schedule of the doctors, 13.8% (n = 13) to receive hygienic and health information, and 2.1% (n = 2) to do research.

All participants had cellphones and could receive and send text messages. They received 1 to 8 messages per day, but only about 10% (n = 21) of them received a maximum of 2 messages a day.

Participants’ Preferences for Notification of Test Results

Notifying the participants that their results are in via text on the same day the laboratory received the results was favored by 98% (n = 196) of participants, and all participants preferred to receive their laboratory test results online. When exploring the possible modes of laboratory result notification (by e-mail or on the patient portal), 52.5% (n = 105) of participants preferred notification by e-mail and 47.5% (n = 95) through the hospital website. About half of the participants (n = 101; 50.55%) believed that viewing laboratory test results online is completely useful. The rest, except 1%, felt this method of notification is partially useful.

Most participants (n = 154; 77%) gave “saving time” as the main reason for receiving laboratory test results electronically. Thirty-one percent (n = 62) said their main reason was preventing the paperwork from becoming lost by the participants (

Regarding the nature of the results, 82.5% of participants (n = 165) preferred to receive both normal and abnormal test results online, while 14% (n = 28) wanted to receive only normal results online and 3.5% (n = 7) only abnormal test results online. Most participants also noted that they would want their online results to be accompanied by such data as an identification number (n = 123; 61.5%) and name and age (n = 109; 54.5%) (Table 3).

Timeliness was the most important characteristic reported by 71.5% (n = 143) of the participants. There were significant relationships between level of education and the reasons given for wanting to receive test results online (

= .0028) and between education level and the data that participants preferred to be reported along with the test results (

= .0175) (Table 3).As noted in Table 3, there were no relationships between other demographic variables and variables.

Participants’ Trust in the Confidentiality of the Online Delivery Methods

From the participants’ perspective, the most reliable method to retain confidentiality was receiving the test results via e-mail (n = 82; 41%). However, about one-sixth of the participants (n = 32; 16%) believed that neither of the suggested methods—e-mail or website&mdash;for sending the test results electronically is secure enough to maintain confidentiality (Table 3).

A significant relationship existed between the frequency of internet use and the patients' trust in the confidentiality of the delivery methods (

 <.0001); frequent internet users most mistrusted the confidentiality. Also, the relationship between the reasons for wishing to get lab test results through online services and participant trust in the confidential method was significant (

= .0253). There was no relationship between demographic variables and the participants’ trust in the confidentiality of the delivery online methods (

To our knowledge, there is no previous study in Iran that has investigated patient preferences regarding the delivery of test results. Most studies done in other countries have evaluated the perspective of special groups of patients, such as those with HIV

 or those with sexually transmitted diseases.

 Some other studies have evaluated security and confidentiality factors into preferences

 and patient priorities for sending normal test results.

 Our study investigated the preferences of a general population of patients who had been referred to a hospital laboratory, regarding the results of all tests.

The findings of our study show that the participants believed that the most useful method for communicating laboratory test results was providing the results to those with internet access through e-mail or website services. Frequent internet users had less confidence in the process of receiving test results via a website. However, although the result was not significant, patients expressed that sending the test results via e-mail was the most satisfactory and confidential notification method.

The findings of a study done by Platteau and colleagues

 that questioned 898 men from Belgium who have sex with men showed a high level of satisfaction among the participants for receiving their HIV test results online. Findings by Shirts and colleagues

 about 96 US participants (48 physicians and 48 nurse practitioners working in the nursing home setting) also indicated that they would be willing to try a Web-based system to get their patients’ test results if they were convinced of its security and confidentiality. The results of another study done in South Carolina and Mississippi in 2009 and 2010 also indicated that this method of communicating test results would accelerate the treatment process.

In other available research, the most important features of the communicating system for test results, from the perspective of the participants, were the timeliness and confidentiality of the data. The results of interviews with 20 Colorado patients, and of a study of 171 women from the United Kingdom with breast cancer who were asked about methods of receiving breast biopsy results, also show that when receiving test results electronically, their timely receipt is highly important to the patients.

In other studies, respondents noted that receiving detailed information along with test results online was important when receiving Pap smear results and for gastroenterologists who were communicating endoscopy findings.

 Providing identification information with their test results was also favored by our study participants, likely because they want assurance that they received the correct test results. This finding suggests that although patients worry about confidentiality, they also consider potential mishandling of the information.

Healthcare institutions should consider their patients’ confidentiality concerns and be sure an effective and efficient identification plan is established and maintained in all patient communication systems. An admission code is also among the data that patients would request, perhaps because they believe that this code gives them the chance to track their results more effectively.

In the Colorado study by Baldwin and colleagues,

 the 20 patients were less enthused about getting test results online because of security concerns. Only 10% of those patients chose e-mail for receiving their results, which is not in line with the current findings. This difference might be because in the current study, when patients asked about confidentiality issues, the researcher explained the details of the approach the laboratory department intended to use in e-mailing test results. The patients were told they would be issued an ID number when initially taking the test and that individualized number would need to be entered to see the test results. With such an approach, unauthorized persons cannot easily access the data. However, the acceptability and feasibility of this approach should be studied.

 on 50 HIV-infected patients from rural Uganda showed that although communication of information about HIV laboratory test results via text messages would be acceptable, a few patients also expressed concerns about confidentiality, mentioning they might be more comfortable with both coded messages and PIN codes to protect confidentiality.

The results of our study confirmed the results of others that indicated patients’ tendency to wish to receive their normal and abnormal test results through different technologies.

The main reason given for receiving results online was time savings, reported by 77% of participants, followed by lowering the chance of missing the results (31%).

Patients' Preferences for Receiving Laboratory Test Results

This study examined the relationship between medication adherence, cost sharing measured as out-of-pocket spending, and total annual spending in Medicare beneficiaries with type 2 diabetes (T2D) to evaluate whether pharmacy cost-sharing programs have the potential to decrease adherence. These programs may unintentionally increase the risk of medical complications and may result in higher spending overall.

This retrospective study used 2006 to 2009 Medicare claims data. The sample included patients 65 years or older with T2D (at least 1 claim with International Classification of Diseases, 9th Revision, Clinical Modification codes 250.x0 and 250.x2 and at least 1 antidiabetes drug claim). 

Medication adherence was measured as proportion of days covered over the first 12 months of observation. Spending and adherence outcomes were defined in deciles. 

The sample included 12,305 patient-year observations. Pharmacy spending for patients in the most adherent (10th) decile was 59% higher than that for patients in the least adherent (1st) decile ($4839 vs $3046). Yet, patients in the 10th decile had 49% lower total ($12,531 vs $24,468) and 64% lower medical spending ($7692 vs $21,421) than patients in the 1st decile. Greater out-of-pocket spending was correlated with lower adherence and higher total and medical spending. 

 This study describes a widespread variation in medication adherence, pharmacy cost sharing, and medical spending in a sample of Medicare beneficiaries with T2D. We found that lower adherence was correlated with higher cost sharing in the Medicare population, perhaps because of unobserved confounding factors. However, the existing literature on patients with employer-sponsored insurance suggests some of this correlation may be indicative of causal relationships.

		Increased cost sharing could decrease the likelihood that patients with type 2 diabetes will adhere to their medication and increase diabetes complications and total diabetes spending. 

		Using Medicare claims data (2006-2009), we examined the relationships between medication adherence, cost sharing measured as out-of-pocket spending, and total annual spending to evaluate whether pharmacy cost-sharing programs have the potential to decrease adherence. 

		We found a widespread variation in medication adherence, pharmacy cost sharing, and medical spending, and that lower adherence is correlated with higher cost sharing in the sample of Medicare beneficiaries with type 2 diabetes.

In 2014, nearly 30 million Americans (9.3%) were living with diabetes, including an estimated 11.2 million who were 65 years or older.

 In 2012, the total national spending on diabetes was $245 billion, including $176 billion in direct medical spending. Patients 65 years or older use more healthcare services than younger patients and incur the majority of costs associated with diabetes. Ultimately, the Medicare program—in which 28% of beneficiaries have diabetes&mdash;bears most of this burden.

 Interventions targeted at individuals using the most resources and who have the greatest expenditures have the potential for a disproportionate impact on spending.

Several studies evaluating the economic burden of diabetes have estimated that antidiabetes agents and supplies account for 12% of total healthcare spending.

 Payers have implemented various cost-sharing schemes to reduce pharmacy spending, including incentive-based formularies, coinsurance or co-payments, and benefit caps.

 Research consistently shows that higher co-payments or coinsurance reduce pharmaceutical utilization and adherence among patients with chronic diseases like type 2 diabetes (T2D), heart failure, and schizophrenia,

 and that reducing co-payments in employer-sponsored insurance plans can increase adherence.

This study describes a widespread variation in medication adherence, pharmacy cost sharing, and medical spending. Increased cost sharing may decrease adherence and increase total diabetes spending.

April 2017

Can Primary Care Physicians Accurately Predict the Likelihood of Hospitalization in Their Patients?

The Breathmobile Improves the Asthma Medication Ratio and Decreases Emergency Department Utilization

Patients' Preferences for Receiving Laboratory Test Results

The Relationship Between Adherence and Total Spending Among Medicare Beneficiaries With Type 2 Diabetes

Medication Burden in Patients With Acute Coronary Syndromes

Concentration of High-Cost Patients in Hospitals and Markets

Provider Type and Management of Common Visits in Primary Care

Physician Variation in Lung Cancer Treatment at the End of Life

Battling the Chargemaster: A Simple Remedy to Balance Billing for Unavoidable Out-of-Network Care

Real-World Evidence and the Behavioral Economics of Physician Prescribing