Instrumental variables regression analysis indicated that inpatient oral nutritional supplement use decreased length of stay, episode cost, and 30-day readmission probability.
To assess the effect of inpatient oral nutritional supplement (ONS) use on length of stay, episode cost, and 30-day readmission probability.
Eleven-year retrospective study (2000 to 2010).
Analyses were conducted using the Premier Perspectives Database, which contained information on 44.0 million adult inpatient episodes. Using a matched sample of ONS and non-ONS episodes for any inpatient diagnosis, instrumental variables regression analysis was performed to quantify the effect of ONS use on length of stay, episode cost, and probability of approximate 30-day readmission. For the readmission outcome, the matched sample was restricted to episodes where the patient was known to be at risk of readmission. The fraction of a hospital’s episodes in a given quarter involving ONS was used as an instrumental variable.
Within the database, 1.6% of 44.0 million adult inpatient episodes involved ONS use. Based on a matched sample of 1.2 million episodes, ONS patients had a shorter length of stay by 2.3 days (95% confidence interval [CI] — 2.42 to –2.16), from 10.9 to 8.6 days (21.0% decline), and decreased episode cost of $4734 (95% CI – $4754 to – $4714), from $21,950 to $17,216 (21.6% decline). Restricting the matched sample to the 862,960 episodes where patients were readmitted at some point, ONS patients had a reduced probability of early readmission (within 30 days) of 2.3 percentage points (95% CI – 0.027 to – 0.019), from 34.3% to 32.0% (6.7% decline).
Use of ONS decreases length of stay, episode cost, and 30-day readmission risk in the inpatient population.
(Am J Manag Care. 2013;19(2):121-128)Malnutrition is a serious but underappreciated problem among hospitalized patients. There is relatively little evidence evaluating the large-scale effectiveness of therapeutic interventions against malnutrition. We conducted an instrumental variables analysis to determine the effect of oral nutritional supplement use in the inpatient setting.
Malnutrition is a serious and underappreciated problem among hospitalized patients. Malnourished patients face heightened risks of poor outcomes, including increased length of stay (LOS),1-3 healthcare costs,1-4 complication rates,2,4-7 readmission rates,8,9 and mortality.2,10-12
Estimates of malnutrition prevalence in the inpatient population range from 8% to 62%, depending on the location and the specific patient population considered.13-17 Groups at highest risk include elderly as well as oncology and gastroenterology patients.16 Despite evidence documenting the deleterious effects of malnutrition in the inpatient setting, studies suggest it is a common problem that often goes unrecognized and undertreated.15,18
A growing body of evidence suggests that oral nutrition supplements (ONS), which deliver both macronutrients and micronutrients for special medical purposes in addition to normal food, might improve outcomes among hospitalized patients. A variety of benefits have been found for ONS use, including reduced LOS,3 inpatient episode cost,3,19 complication rates,19,20 depressive symptoms,21 and readmission rates,22,23 and improved lean body mass recovery.24 However, previous studies suffer from limitations, including modest sample sizes, narrowly selected patient populations, and in observational studies, possible selection bias. Consequently, questions remain regarding the robustness and generalizability of existing findings and the size of gains and healthcare costs associated with ONS use in hospitalized patients.
This retrospective data analysis was conducted to assess the association and causal impacts of ONS on health outcomes for hospitalized patients, focusing on 3 key outcomes: LOS, episode cost, and probability of 30-day readmission.
Setting, Subjects, and Data Sources
This analysis was conducted using the Premier Perspectives Database. This database contains diagnostic and billing information on 44.0 million adult inpatient episodes at 460 sites during the years 2000 to 2010. Premier estimates that these data cover 20% of all US inpatient episodes. The sample was restricted to adults 18 years or older and excluded terminal episodes and all episodes involving tube feeding, leaving only oral feeding for examination. All monetary figures were reported in 2010 dollars and inflation-adjusted based on the Bureau of Labor Statistics medical Consumer Price Index (http://www.bls.gov/ cpi/#tables).
The study’s 3 key outcome variables were LOS, episode cost, and probability of 30-day readmission. Length of stay was defined as the number of days of direct patient care (minimum 1 day) from admission to discharge. Episode cost was defined as the actual costs to treat the patient during the hospitalization. Thirty-day readmission probability was defined as a return hospitalization for any diagnosis. For patient confidentiality purposes, the Premier database only contains the month and year of an inpatient episode. Therefore, the 30-day readmission window was calculated by identifying admissions later the same month or during the following month. Given that there are no International Classification of Diseases, Ninth Revision or Current Procedural Terminology codes that identify ONS use, ONS was defined as a “complete nutritional supplement, oral,” as indicated by the Premier data, and coded as a binary variable, indicating any ONS used during the inpatient episode.
Naïve ordinary least squares (OLS) regression analyses were performed on the full matched sample. Analyses controlled for a variety of patient, episode, and provider characteristics. Demographic covariates included age, age squared, insurance type, marital status, race, and sex. Comorbidity covariates included all components of the Charlson Comorbidity Index.25,26 Health history covariates included whether the patient had been admitted to any Premier network hospital in the previous 6 months, and whether the patient was admitted from the emergency department, by physician referral, or by inter-facility transfer. Hospital- specific covariates included number of beds, urban location, whether the site was a teaching hospital, and region (Northeast, Midwest, West, or South as defined by US Census data). Time trends were controlled for using year and quarter dummies.
Treatment is assigned randomly in clinical trials to avoid confounding; however, there is potential for selection bias in observational research from unobserved factors that may influence study outcomes.27,28 Because it is likely to be administered to individuals who are less healthy, ONS use could be spuriously associated with increased LOS. Additionally, only certain patient health-related risk factors were directly observable with available data. Therefore, further methods were used to remove these sources of potential selection bias: propensity score matching and instrumental variables analysis.
Propensity Score Matching. To diminish the potential for confounding due to differences in observed personal characteristics and to identify nutritionally at-risk patients, propensity score matching29 was used to match ONS episodes to similar non-ONS episodes. The probability of receiving ONS was estimated using a logistic regression of ONS based on the covariates noted above. After removing all episodes involving children (<18 years) and tube feeding, each ONS episode was matched to its nearest non-ONS episode neighbor.
Instrumental Variables Analysis. Instrumental variable analysis was used to specifically address potential bias due to nonrandomized treatment selection, which could not be addressed with propensity matching alone. Instrumental variables can remove the effect of selection bias and identify the causal effect of a treatment on outcomes.30-32 Using this method requires an instrument that correlates with the treatment of interest but does not affect the outcome, except through its influence on the likelihood of receiving treatment.
For this analysis, the selected instrument was the fraction of episodes involving any ONS use in a given hospital in a given quarter. By looking at changes in ONS use based on a hospital’s inclination to prescribe it, rather than underlying patient characteristics, the unbiased identification of the effect of ONS was made feasible. Since instrumental variable properties are best understood in linear settings,33 this instrument was applied to linear models of the 3 outcomes. Several tests of the instrument’s validity were performed.
To control further for unobserved patient heterogeneity, the model included fixed effects for groups based on how long patient data remained observable prior to loss to follow-up. These fixed-effects “follow-up” groups were no patient follow- up data; 1 day through 1 year of follow-up; 1 to 2 years of follow-up; 2 to 3 years of follow-up; and more than 3 years of follow-up. Because life expectancy cannot generally be observed in the Premier database (except when individuals die in a Premier network hospital), follow-up duration served as a proxy for underlying health status. Observed follow-up using hospital-based data may be a preferable measure of overall patient frailty, because the diagnostic codes present in a single episode (vs multiple follow-up episodes) are unlikely to reflect the full range of patient comorbidities.
Additional Modeling of Readmissions. For the readmission outcome, the matched sample was restricted to episodes where the patient was known to be at risk of readmission following discharge. If patients did not die following hospitalization, it could be assumed ONS had 2 potential benefits: it prevented readmissions by making people healthy, or it delayed readmissions among those eventually readmitted. Because the Premier data did not distinguish between patients not readmitted due to recovery and those not readmitted due to death, the current study could only measure the effect of delayed readmission (by calculating the change in 30-day readmission probability among patients eventually readmitted). This approach also provided a conservative estimate of the total impact of ONS on readmission.
Return on Investment Calculations. Next, estimates for the effect of ONS on LOS, episode cost, and readmission probability were used to calculate a return on investment (ROI) for ONS use, using the following formula:
ROI = savings generated through ONS use — amount spent on ONS
amount spent on ONS
The above formula yields the “episode cost” ROI of ONS use through reduced episode cost. Savings generated from ONS use were defined as the average reduction in episode cost due to ONS use. The amount spent on ONS was the average episode cost of ONS use. For readmission ROI, the average episode cost among the readmitted population was multiplied by the reduction in the probability of readmission to calculate the savings generated through ONS use. The amount spent on ONS was defined as for the episode cost ROI.
Computation. Analyses were performed using Stata version 11 (StataCorp LP, College Station, Texas). A 2-sided P value of .05 or less was considered statistically significant. A detailed summary of additional testing and sensitivity analyses conducted to validate study results can be found in the (available at www.ajmc.com).
From 46.1 million inpatient episodes and 810,589 episodes involving ONS use, we excluded 306,528 tube-feeding episodes, 1,798,907 involving patients under age 18 years, 112 with incomplete data, and 19,817 terminal episodes to obtain a sample of 44.0 million episodes and 724,027 ONS episodes. The overall rate of ONS use in adult inpatient episodes was 1.6%. Each adult ONS episode was matched to an adult non-ONS episode, to obtain a matched sample of 1,160,088 episodes.
Mean characteristics of ONS episodes, all non-ONS episodes, and matched non-ONS episodes are reported in . Compared with general non-ONS inpatient episodes, individuals receiving ONS were older (age 68.4 vs 56.7 years) and less healthy on various dimensions; and 42.2% of ONS episodes were preceded by an admission in the prior 6 months, compared with only 25.6% for non-ONS episodes. The average LOS for an ONS episode was 12.5 days compared with 4.8 days for non-ONS episodes.
shows the characteristics of matched ONS sample subgroups by follow-up group. Patient ONS use was highly correlated with other health markers, including prior admission history, LOS, episode cost, and Charlson Comorbidity Index score.
Length of Stay
Ordinary least squares regression analysis performed on the full matched sample showed that ONS use was associated with a 2.9-day (95% confidence interval [CI] 2.8-3.0 days), or 34.7%, increase in LOS, from 8.3 to 11.2 days. However, when instrumental variables regression analysis was used to account for selection bias, ONS lowered LOS by 2.3 days (95% CI -2.4 to -2.2 days), or 21.0%, from 10.9 to 8.6 days (, columns 1 and 2).
Next, to determine whether the effect of ONS differed depending on the underlying health status of the treated individual, the instrumental variables regression analysis was repeated on matched sample subgroups, sorted by duration of observed follow-up. For this comparison, all episodes with no observed follow-up were dropped. Once the matched sample was restricted in this way (Table 3, columns 3-6), episodes with longer follow-up duration were successively dropped to create an increasingly sick sample moving from column 3 (patients with at least 1 day of follow-up) to column 6 (patients with 1 day to 1 year of follow-up). When the data were grouped in this way, it became apparent that ONS had the greatest LOS benefit for the sickest group (-22.8%) and a smaller, but still significant, benefit for the healthiest group (-16.3%).
Ordinary least squares regression analysis showed that ONS use was associated with an increased episode cost of $7598 (95% CI $7579-$7617), or 50.7%, from $14,998 to $22,596 (, column 1). However, when the instrumental variables method was applied to the full matched sample (Table 4, column 2), ONS use decreased episode cost by $4734 (95% CI -$4754 to -$4714), or 21.6%, from $21,950 to $17,216. When the matched samples were grouped in terms of duration of known follow-up (Table 4, columns 3-6), a clear pattern was observed, with the largest ONS benefit going to the sickest individuals. Episode cost savings ranged from 17.9% to 24.0% for the healthiest to the sickest subgroups, respectively.
In the known follow-up subsample, naïve OLS regressions showed that ONS use was associated with a 0.3 percentage point (95% CI -0.005 to -0.001), or 0.9%, decrease in readmission probability, from 33.4% to 33.1% (, column 1). Instrumental variables regression results demonstrated that ONS use led to a 2.3 percentage point (95% CI −0.027 to −0.019), or 6.7%, decrease in the probability of readmission among episodes with any follow-up, from 34.3% to 32.0% (Table 5, column 2). Assuming conservatively that ONS provided no benefit to patients never readmitted and served only to delay readmissions among those who were eventually readmitted, this finding implied that ONS decreased the probability of readmission in the full matched sample by at least 6.9% (measured as a 0.0231 reduction in readmission probability multiplied by the 74% of the matched sample eventually readmitted, divided by a baseline 30-day readmission rate of 24.7% in the matched sample). Grouping the subsample with known follow-up by underlying health status (Table 5, columns 2-5) again shows a clear pattern of the largest benefit of ONS use going to the sickest individuals (14.1%).
Return on Investment
Use of ONS cost an average of $88.26 per episode. This cost included the cost of ONS and associated labor and administrative expenses, based on hospital reporting. When held against the estimate (Table 4, Column 2) that ONS use generates $4734 in savings per episode, this amounted to an ROI of $52.63 in net savings for every dollar spent on ONS in terms of reduced episode cost.
To calculate readmission ROI, it was assumed that hospital sites could not distinguish between individuals who would eventually be readmitted and those who would not, and therefore had to administer ONS to all matched sample patients. As noted previously, study estimates indicated that ONS decreased readmission probability by 0.0231 for the 74% of the matched sample eventually readmitted. This conservatively assumed no benefit from readmission prevention for the other 26%. This effect was then multiplied by $18,478 (the average population episode cost for inpatient readmission), resulting in an estimated $314.13 in savings per episode due to ONS use. This translated into an ROI of at least $2.56 in net savings due to averted 30-day readmissions for every dollar spent on ONS in the matched sample.
This study found that ONS use in hospitalized patients led to substantial reductions in LOS, episode cost, and 30- day readmissions. Specifically, ONS use resulted in a 2.3-day (21.0%) LOS decrease, $4734 (21.6%) in decreased episode costs, and a 6.7% decrease in 30-day readmissions among patients eventually readmitted. Conservatively assuming no benefit to those never readmitted, these outcomes translated to a minimum 6.9% decrease in readmissions among the full matched sample of all ONS episodes and similar non-ONS episodes. The study of 30-day readmissions is particularly relevant, given new Medicare rules that may make hospitals liable for some readmissions within 30 days.34-40
These gains, it is important to note, are consistent with results from previous randomized controlled trials. In a study of general inpatients, Somanchi and colleagues found that early nutritional intervention reduced LOS by 1.93 days (P = .003), and in a severely malnourished subpopulation, reduced LOS by 3.2 days (P = .052).3 In a UK-based study, Lawson and colleagues found that ONS was associated with a 6% reduction in episode cost.19 Somanchi et al found a $1514 episode cost decrease among severely malnourished patients.3 This cost reduction was lower than that observed in the current study. However, Somanchi et al calculated cost savings as number of days of reduced LOS multiplied by average cost of additional days. This approach did not take into account that ONS use might make the inpatient stay less resource intensive, not just shorter.
In a trial with malnourished patients, Norman and colleagues found that ONS use decreased 3-month readmissions from 48% to 26%.23 Likewise, Gariballa and colleagues found that ONS use led to a 28% reduction in 6-month readmissions, from 40% to 29% (adjusted hazard ratio 0.68 [95% CI = 0.49-0.94]).22 However, in both of these randomized controlled trials, ONS use was sustained postdischarge. For the current study, it was not possible to determine whether patients continued ONS after leaving the hospital.
Because ONS is inexpensive to provide, the sizable savings generated make it a cost-effective therapy. From the healthcare perspective, for every dollar spent on ONS, the ROI was $52.63 in immediate net episode cost savings and $2.56 in net savings from avoided 30-day readmissions. The 1:1 matched sample estimates imply that doubling ONS use by targeting patients similar to current ONS users is likely to produce financial returns to hospitals and improve patient outcomes. Sensitivity analyses suggest that further increases beyond doubling may continue to generate positive results, but more research is needed on this point.
The current study has 2 key advantages over previous research. First, it used a large database to estimate the effect of ONS based on real-world data. With 44 million adult inpatient episodes, these data were relevant and broadly representative. Second, econometric methods were used to enable causal inference regarding the impact of ONS on patient outcomes. By applying propensity score matching and instrumental variables, potential bias due to nonrandom selection into ONS treatment was mitigated. This made it possible to estimate causal impact of ONS use on LOS, episode cost, and readmission probability.
However, the Premier Perspectives data did have limitations. The lack of detailed patient health information, such as laboratory test results and patient health status assessment, led to a selection challenge whereby patients receiving ONS were presumably sicker on a variety of dimensions not fully observable in the data. This limitation was addressed using propensity score matching and instrumental variables analysis. In addition, the fact that it was not possible to distinguish between avoided readmissions due to recovery, death, or transfer to a non-Premier hospital meant that analyses of the effect of ONS on readmission had to be confined to a subsample of episodes with known follow-up. Therefore, the benefit of ONS could only be quantified based on delayed, rather than prevented, readmission. The Premier data set did not provide data on ONS use following discharge. Lastly, although we performed multiple instrument validity tests, more comprehensive tests could be performed with hospital-specific quality measures such as report cards. In the future, researchers with access to more comprehensive data may be able to gain additional insight on this issue.
Using the instrumental variables method, this study found that the use of ONS led to statistically significant decreases in inpatient LOS, episode cost, and readmission. Given the high prevalence of malnutrition among inpatient populations, these results suggest that ONS use could help improve outcomes at relatively low cost to the healthcare system. Today, hospitals are facing pressures to find low-cost, highly effective therapy while maintaining quality of care. By increasing ONS use, hospitals can improve hospitalization outcomes and decrease healthcare spending.
Author Affiliations: From University of Chicago (TJP), Chicago, IL; Precision Health Economics, (JTS, BS), Los Angeles, CA; University of Southern California (DNL, DPG), Los Angeles, CA.
Funding Source: Abbott Nutrition.
Author Disclosures: Dr Snider and Mr Stryckman report involvement in the preparation of this manuscript while employed by Precision Health Economics, which received consulting fees from Abbott Nutrition. Drs Philipson, Lakdawalla, and Goldman are partners at Precision Health Economics.
Authorship Information: Concept and design (JTS, DNL, DPG, BS); acquisition of data (BS); analysis and interpretation of data (TJP, JTS, DNL, DPG); drafting of the manuscript (TJP, JTS, DNL, DPG, BS); critical revision of the manuscript for important intellectual content (TJP, JTS, DNL, DPG, BS); statistical analysis (JTS, BS); provision of study materials or patients (JTS, BS); obtaining funding (TJP, JTS, DNL, DPG, BS); and supervision (TJP, JTS, DNL, DPG, BS).
Address correspondence to: Julia Thornton Snider, PhD, Precision Health Economics, 11100 Santa Monica Blvd, Ste 500, Los Angeles, CA 90025. E-mail: email@example.com.
1. Chima CS, Barco K, Dewitt ML, Maeda M, Teran JC, Mullen KD. Relationship of nutritional status to length of stay, hospital costs, and discharge status of patients hospitalized in the medicine service. J Am Diet Assoc. 1997;97(9):975-978.
2. Correia MI, Waitzberg DL. The impact of malnutrition on morbidity, mortality, length of hospital stay and costs evaluated through a multivariate model analysis. Clin Nutr. 2003;22(3):235-239.
3. Somanchi M, Tao X, Mullin GE. The facilitated early enteral and dietary management effectiveness trial in hospitalized patients with malnutrition. JPEN J Parenter Enteral Nutr. 2011;35(2):209-216.
4. Braunschweig C, Gomez S, Sheean PM. Impact of declines in nutritional status on outcomes in adult patients hospitalized for more than 7 days. J Am Diet Assoc. 2000;100(11):1316-1322.
5. Schoonhoven L, Grobbee DE, Donders AR, et al; prePURSE Study Group. Prediction of pressure ulcer development in hospitalized patients: a tool for risk assessment. Qual Saf Health Care. 2006;15(1): 65-70.
6. Fry DE, Pine M, Jones BL, Meimban RJ. Patient characteristics and the occurrence of never events. Arch Surg. 2010;145(2):148-151.
7. Schneider SM, Veyres P, Pivot X, et al. Malnutrition is an independent factor associated with nosocomial infections. Br J Nutr. 2004;92(1): 105-111.
8. Allaudeen N, Vidyarthi A, Maselli J, Auerbach A. Redefining readmission risk factors for general medicine patients. J Hosp Med. 2011;6(2):54-60.
9. Mudge AM, Kasper K, Clair A, et al. Recurrent readmissions in medical patients: a prospective study. J Hosp Med. 2011;6(2):61-67.
10. Gariballa S, Forster S. Malnutrition is an independent predictor of 1-year mortality following acute illness. Br J Nutr. 2007;98(2):332-336.
11. Sullivan DH, Walls RC. Protein-energy undernutrition and the risk of mortality within six years of hospital discharge. J Am Coll Nutr. 1998;17(6):571-578.
12. Sullivan DH, Sun S, Walls RC. Protein-energy undernutrition among elderly hospitalized patients: a prospective study. JAMA. 1999;281(21): 2013-2019.
13. Bauer JD, Isenring E, Torma J, Horsley P, Martineau J. Nutritional status of patients who have fallen in an acute care setting. J Hum Nutr Diet. 2007;20(6):558-564.
14. Naber TH, Schermer T, de Bree A, et al. Prevalence of malnutrition in nonsurgical hospitalized patients and its association with disease complications. Am J Clin Nutr. 1997;66(5):1232-1239.
15. McWhirter JP, Pennington CR. Incidence and recognition of malnutrition in hospital. BMJ. 1994;308(6934):945-948.
16. Pirlich M, Schütz T, Norman K, et al. The German hospital malnutrition study. Clin Nutr. 2006;25(4):563-572.
17. Planas M, Audivert S, Pérez-Portabella C, et al. Nutritional status among adult patients admitted to an university-affiliated hospital in Spain at the time of genoma. Clin Nutr. 2004;23(5):1016-1024.
18. Lyder CH, Preston J, Grady JN, et al. Quality of care for hospitalized medicare patients at risk for pressure ulcers. Arch Intern Med. 2001;161(12):1549-1554.
19. Lawson RM, Doshi MK, Barton JR, Cobden I. The effect of unselected post-operative nutritional supplementation on nutritional status and clinical outcome of orthopaedic patients. Clin Nutr. 2003;22(1): 39-46.
20. Beattie AH, Prach AT, Baxter JP, Pennington CR. A randomised controlled trial evaluating the use of enteral nutritional supplements postoperatively in malnourished surgical patients. Gut. 2000;46(6):813-818.
21. Gariballa S, Forster S. Effects of dietary supplements on depressive symptoms in older patients: a randomised double-blind placebo-controlled trial. Clin Nutr. 2007;26(5):545-551.
22. Gariballa S, Forster S, Walters S, Powers H. A randomized, doubleblind, placebo-controlled trial of nutritional supplementation during acute illness. Am J Med. 2006;119(8):693-699.
23. Norman K, Kirchner H, Freudenreich M, Ockenga J, Lochs H, Pirlich M. Three month intervention with protein and energy rich supplements improve muscle function and quality of life in malnourished patients with non-neoplastic gastrointestinal disease—a randomized controlled trial. Clin Nutr. 2008;27(1):48-56.
24. Jensen MB, Hessov I. Dietary supplementation at home improves the regain of lean body mass after surgery. Nutrition.1997;13(5):422-430.
25. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40(5):373-383.
26. Charlson ME, Sax FL, MacKenzie CR, Braham RL, Fields SD, Douglas RG Jr. Morbidity during hospitalization: can we predict it? J Chronic Dis. 1987;40(7):705-712.
27. Linden A, Adams JL. Evaluating disease management programme effectiveness: an introduction to instrumental variables. J Eval Clin Pract. 2006;12(2):148-15å4.
28. Goldman DP, Bhattacharya J, McCaffrey DF, et al. Effect of insurance on mortality in an HIV-positive population in care. J Am Stat Assoc. 2001;96(455):883-894.
29. Rubin DB. Estimating causal effects from large data sets using propensity scores. Ann Intern Med. 1997;127(8, pt 2):757-763.
30. Bao Y, Duan N, Fox SA. Is some provider advice on smoking cessation better than no advice? an instrumental variable analysis of the 2001 National Health Interview Survey. Health Serv Res. 2006;41(6): 2114-2135.
31. McClellan M, McNeil BJ, Newhouse JP. Does more intensive treatment
of acute myocardial infarction in the elderly reduce mortality?
analysis using instrumental variables. JAMA. 1994;272(11):859-866.
32. Newhouse JP, McClellan M. Econometrics in outcomes research: the use of instrumental variables. Annu Rev Public Health. 1998; 19:17-34.
33. Staiger D, Stock JH. Instrumental variables regression with weak instruments. Econometrica. 1997;65(3):557-586.
34. Centers for Medicare & Medicaid Services. Medicare Program; Hospital Inpatient Prospective Payment Systems for Acute Care Hospitals and the Long-Term Care Hospital Prospective Payment System and Fiscal Year 2012 Rates; Revisions to the Reductions and Increases to Hospitals’ FTE Resident Caps for Graduate Medical Education Payment. CMS-1518-F. https://www.cms.gov/Medicare/Medicare-Feefor-Service-Payment/AcuteInpatientPPS/FY-2012-IPPS-Final-Rule-Home- Page-Items/CMS1250103.html. Published 2011. Accessed January 7, 2013.
35. Andrews M. Health law forces changes to reduce hospitals readmissions. Kaiser Health News. http://www.kaiserhealthnews.org/features/ insuring-your-health/michelle-andrews-on-hospital-readmissions.aspx?referrer=search. Published February 22, 2011. Accessed January 7, 2013.
36. Centers for Medicare & Medicaid Services. Acute patient PPS: overview. http://www.cms.gov/AcuteInpatientPPS/01_overview.asp. Published 2011. Accessed February 15, 2012.
37. Foster D, Harkness G. Healthcare reform: pending changes to reimbursement for 30-day readmissions. In: Research Brief 2010. Ann Arbor, MI: Thomson Reuters;
38. Gold N. 3 Readmissions to reduce now. http://www.healthleadersmedia. com/page-1/COM-263665/3-Readmissions-to-Reduce-Now. Published March 15, 2011. Accessed February 14, 2012.
39. Stone J, Hoffman GJ. Medicare Hospital Readmissions: Issues, Policy Options and PPACA. Congressional Research Service. http:// www.pascenter.org/publications/publication_home.php?id=1230. Published September 1, 2010. Accessed January 7, 2013.
40. Thorpe JH, Cascio T. Medicare hospital readmissions reduction program. Legal Notes. Volume 3, issue 4. http://www.rwjf.org/en/ research-publications/find-rwjf-research/2011/10/medicare-hospitalreadmissions- reduction-program.htm. Published October 1, 2011. Accessed January 7, 2013.