Demonstrated health care cost savings for women: findings from a community health worker intervention designed to address depression and unmet social needs

To evaluate the impact of a community health worker intervention (CHW) (referred to as Personalized Support for Progress (PSP)) on all-cause health care utilization and cost of care compared with Enhanced Screening and Referral (ESR) among women with depression. A total of 223 patients (111 in PSP and 112 in ESR randomly assigned) from three women’s health clinics with elevated depressive symptoms were enrolled in the study. Their electronic health records were queried to extract all-cause health care encounters along with the corresponding billing information 12 months before and after the intervention, as well as during the first 4-month intervention period. The health care encounters were then grouped into three mutually exclusive categories: high-cost (> US$1000 per encounter), medium-cost (US$201–$999), and low-cost (≤ US$200). A difference-in-difference analysis of mean total charge per patient between PSP and ESR was used to assess cost differences between treatment groups. The results suggest the PSP group was associated with a higher total cost of care at the baseline; taking this baseline difference into account, the PSP group was associated with lower mean total charge amounts (p = 0.008) as well as a reduction in the frequency of high-cost encounters (p < 0.001) relative to the ESR group during the post-intervention period. Patient-centered interventions that address unmet social needs in a high-cost population via CHW may be a cost-effective approach to improve quality of care and patient outcomes.

     

Technology in health care: the social impact and economic cost

The contribution of technology to longevity and the quality of life has been substantial during the twentieth century. In the past two decades, technology employed in the care of hospitalized patients has been responsible for most of the cost by which the medical inflation rate (Medical Economic Index) exceeds the consumer price index. In most instances, the marginal benefit from the incremental cost is too small to be measured. If this viewpoint is correct, and if governments continue to contain medical costs, the future use of technology will be limited to those applications which have a clearly demonstrable marginal benefit associated with their incremental cost.     

Effect of temperature on transport and plating media for enteric pathogens.

The effect of wide variations in incubation temperatures and long periods of incubation on transport and enrichment broths and plating media was determined by exhaustive analysis of 132 diarrheal stools for salmonellae and shigellae. Homogenized stools were streaked onto eosin methylene blue (EMB), Salmonella-Shigella (SS), and xylose lysine deoxycholate (XLD) agar plates, and into saline, Cary-Blair (CB) transport medium, and Selenite F and gram-negative (GN) enrichment broths. Incubation temperatures were compared at 20 C, 35 C, 40 C and ambient, and over a range of 4 to 52 C for media incubated in an insulated picnic cooler in an auto trunk. At 1, 2, 4, and 7 days the plates were observed, and the broths were subcultured. Each stool was streaked to 12 plates for 48 observations and pickings, and to 48 tubes, subcultured to 192 plates, for a total of 240 observations for pathogens. Analysis of data from 6,246 Salmonella-positive plates showed direct streaking to be most effective after 2 days of incubation, but broths were equally effective at 1 or 2 days. By day 4 many plates were overgrown, and both plates and broths showed diminution of positives by about 10% and at day 7, 19%. The 2,434 Shigella-positive plates were more demanding in all times and temperatures of incubation than salmonellae. Although at day 2 best results were obtained on direct streaking, shigellae die-offs in broths were excessive, with positive declining 23.7% by day 2, 49% by day 4, and 60% by day 7. Direct plating of both pathogens was poor at 20 C with about 48% success, but salmonellae preferred higher temperatures (35 and 40 C), whereas shigellae chose 35 C and ambient, which averaged 28 C for the 10-month study. Temperature was immaterial to salmonellae in broths with ambient slightly better than 35 C, but shigellae preferred 20 C and showed a 50% failure rate at 40 C, ambient being equal to 35 C. The preferential rank of broths in efficacy was GN greater than selenite greater than saline greater than CB greater than direct for salmonellae; for shigellae, GN greater than saline greater than direct greater than CB greater than selenite, with selenite proving to be unsuitable for shigellae. Plating media preferences were XLD greater than EMB greater than SS. Ten of 39 shigellae strains could not be recovered from the selenite and SS media combination, the many replications notwithstanding. The effectiveness of salmonellae and shigellae detection at ambient temperatures in Louisiana during the 10-month study period, as compared to controlled incubation temperatures, indicates that satisfactory enteric bacteriology can be done in warm climates without constant temperature incubators.     

Application of rejection criteria for stool cultures for bacterial enteric pathogens.

Over 20 months, we prospectively assessed the yield of cultures for Salmonella, Shigella, and Campylobacter spp. from adults and children. In the first 10 months, 53% (567 of 1,097) of specimens from adults were from patients who had been in the hospital for > 3 days. Overall, only 2.6% (29 of 1,097) of specimens contained pathogens, and all were from patients who had been in the hospital for < or = 3 days. Over the second 10 months, specimens from adults in the hospital for > 3 days were not cultured unless special reasons existed. Thirty percent (368 of 1,229) of specimens were rejected. Concurrently, 7.5% (51 of 677) of stool specimens from children were positive. Only one positive specimen came from a child who had been in the hospital for > 3 days. Neither stool consistency nor fecal leukocytes were useful predictors for the presence of an enteric pathogen. We recommend that specimens from both adults and children in the hospital for > 3 days not be cultured unless there are delineated plausible clinical or epidemiological reasons to do so.     

Motility-indole-lysine medium for presumptive identification of enteric pathogens of Enterobacteriaceae.

Detection of lysine decarboxylase activity is a useful supplement to reactions on triple sugar-iron (TSI) and urea agars in the initial examination of suspected pathogenic isolates from fecal cultures. Owing to the added value of motility and indole production in the differentiation of enteric pathogens, we prepared and evaluated a motility-indole-lysine (MIL) medium. The following 890 organisms were tested: 264 Shigella, 2 Edwardsiella, 182 Salmonella enteritidis, 235 S. typhi, 3 Arizona, 32 Yersinia enterocolitica, and 172 other members of the family Enterobacteriaceae. With few exceptions the MIL medium gave the same results as the standard motility, indole, and lysine decarboxylase (Moeller) test media. All discrepancies were with the indole reaction, which was weak in 2 of 67 strains of Escherichia coli and falsely negative in 6 of 32 strains of Y. enterocolitica. When both TSI agar and lysine-iron agar (LIA) slants are used in the evaluation isolates from fecal cultures, detection of H2S is duplicated. Both LIA and MIL medium detect lysine decarboxylase and deaminase activity equally well. Because of its ability to detect motility and indole production, the MIL medium is more useful than LIA when used with TSI agar. The combination of TSI agar, MIL medium, and urea agar enables reliable initial recognition of enteric pathogens of the Enterobacteriaceae.     

Parasitic, bacterial, and viral enteric pathogens associated with diarrhea in the Central African Republic.

A total of 1,197 diarrheic children less than 15 years old were investigated for parasitic, bacterial, and viral enteropathogens from March 1981 through February 1982 in the Central African Republic. One or more pathogens were identified from 49.4% of the patients. Rotavirus was the most frequently identified pathogen among children less than 18 months old. Enteropathogenic Escherichia coli was the second most frequently isolated pathogen (12.1%) in children less than 2 years of age. Campylobacter jejuni was also isolated frequently from diarrheic children less than 5 years of age (10.9%). Entamoeba histolytica was identified in very young children and was found to be the most frequent enteropathogen associated with diarrhea in children over the age of 2 years. Enterotoxigenic Escherichia coli was rarely isolated (ca. 2%). There was a peak in the incidence of rotavirus during the dry season and in the incidence of Campylobacter jejuni during the rainy season.     

Utilization and cost savings of a wide-area computer network for neurosurgical consultation.

BACKGROUND AND OBJECTIVE: Telemedicine systems offer many potential advantages for health care delivery. Most reports have centered on the delivery of primary and medical subspecialty care rather than on its impact on patient care and the potential for cost savings. In 1993, we implemented NeuroLink, a wide-area teleradiology network for delivery of specialty care in neurologic surgery at Allegheny General Hospital (AGH). This study was designed to determine the potential cost savings of such a network. METHODS: We prospectively reviewed 100 consecutive telemedicine neurosurgical consultations from 20 western Pennsylvania community hospitals participating in the NeuroLink network. Data related to referring hospital, diagnosis, disposition of the patient, and mode of transportation were reviewed. To determine the potential cost savings, the differential of hospital-based charges between AGH and western Pennsylvania primary hospitals was calculated based on an average length of stay (LOS), patient bed costs, and transportation charges. RESULTS: Of the 100 patients, 33 did not require transfer to a tertiary facility but were instead managed at the community hospital as a direct result of the remote diagnosis and image review disclosing that neurosurgical procedures or intensive care were not required. Cost analysis, comparing the average LOS at AGH with that of the average community hospital, including transportation, showed savings of $502,638. CONCLUSION: Our neurosurgical wide-area computer network has led to more appropriate transfer of patients to a tertiary facility and significant estimated cost savings.     

Rapid automated method for screening of enteric pathogens from stool specimens.

A total of 800 colonies suggestive of Salmonella, Shigella, or Yersinia species isolated on stool differential agar media were inoculated onto both conventional biochemical test media (triple sugar iron agar, urea agar, and phenylalanine agar) and Entero Pathogen Screen cards of the AutoMicrobic system (Vitek Systems, Inc., Hazelwood, Mo.). Based on the conventional tests, the AutoMicrobic system method yielded the following results: 587 true-negatives, 185 true-positives, 2 false-negatives, and 26 false-positives (sensitivity, 99%; specificity, 96%). Both true-positive and true-negative results were achieved considerably earlier than false results (P less than 0.001). The Entero Pathogen Screen card method is a fast, easy, and sensitive method for screening for Salmonella, Shigella, or Yersinia species. The impossibility of screening for oxidase-positive pathogens is a minor disadvantage of this method.     

Tweedie distributions for fitting semicontinuous health care utilization cost data

Background

The statistical analysis of health care cost data is often problematic because these data are usually non-negative, right-skewed and have excess zeros for non-users. This prevents the use of linear models based on the Gaussian or Gamma distribution. A common way to counter this is the use of Two-part or Tobit models, which makes interpretation of the results more difficult. In this study, I explore a statistical distribution from the Tweedie family of distributions that can simultaneously model the probability of zero outcome, i.e. of being a non-user of health care utilization and continuous costs for users.

Methods

I assess the usefulness of the Tweedie model in a Monte Carlo simulation study that addresses two common situations of low and high correlation of the users and the non-users of health care utilization. Furthermore, I compare the Tweedie model with several other models using a real data set from the RAND health insurance experiment.

Results

I show that the Tweedie distribution fits cost data very well and provides better fit, especially when the number of non-users is low and the correlation between users and non-users is high.

Conclusion

The Tweedie distribution provides an interesting solution to many statistical problems in health economic analyses.

     

Cost and time savings following introduction of rejection criteria for clinical specimens.

We have evaluated the yield of several tests and have instituted specimen rejection criteria to reduce costs and save time. For a 12-month period, we recorded the reduction of these tests and calculated the resultant cost and time savings. Seven changes were analyzed: not performing fungal or mycobacterial (acid-fast bacillus) cultures on cerebrospinal fluid (CSF) specimens from patients without known immunosuppression when chemistry and cell count are normal; not performing routine stool culture or ovum and parasite examination on specimens from patients in the hospital for > 3 days; not culturing endotracheal suction aspirates when no organisms or > 10 squamous epithelial cells are present; discontinuing broth cultures on all specimens except for tissue, continuous ambulatory peritoneal dialysis fluid, and CSF from patients with shunts; and eliminating bacterial antigen tests. For each test, the number not performed (n), reagent savings, and technologist time saved, respectively, were as follows: CSF fungal culture, 267, $999, and 67 h; CSF acid-fast bacillus culture, 275, $1,662, and 124 h; stool cultures, 320, $2,991, and 98 h; ovum and parasite examinations, 216, $525, and 108 h; endotracheal suction aspirate cultures, 1,505, $4,447, and 306 h; broth cultures, 5,218, $4,931, and 80 h; and bacterial antigen tests, 2,598, $2,293, and 299 h. Overall, 5,181 tests were rejected and 5,218 broth cultures were omitted. Achievable savings were $28,000 in reagent costs and 1,082 h of technologist time. In conclusion, rejecting specimens of proven low yield saves reagent costs and, more importantly, saves technologist time. This time can be spent on specimens having greater diagnostic utility.     

A centralized approach to out‐of‐province genetic testing leads to cost savings: the Alberta experience

The Genetic Resource Center (GRC) is a centralized process for requesting genetic testing that is not available within the province (Alberta, Canada). In order to assess potential cost savings associated with this process, all applications received by the GRC in 2010 were reviewed, and cost savings were recorded for statistical analysis. Seven areas of cost savings were identified: (i) negotiated pricing, (ii) laboratory selection, (iii) testing setup in‐province, (iv) duplicate testing, (v) inappropriate testing, (vi) sequential testing and (vii) testing offered within the province.The total test cost of the 615 applications submitted in 2010 without the GRC process would have been $766,783 (Canadian dollars). A total cost savings of $112,201 was achieved through the GRC, which represents 15% of the total cost of requested testing ($112,201/$766,783). This is the first study to examine areas of cost savings for genetic testing sent out‐of‐province. The greatest cost savings resulted from the areas of laboratory selection and negotiated pricing. A centralized process to manage out‐of‐province genetic test requests results in consistency in testing and significant cost savings.